When ChatGPT burst onto the scene in late 2022, investors did not waste time picking which tech companies they thought would be winners from the new artificial intelligence (AI) trend. At first, the market went shopping for pick-and-shovel companies. Nvidia (NASDAQ: NVDA) sells the graphics processing units (GPUs) that provide the computing power to train and run generative models. Micron Technology (NASDAQ: MU) sells the high bandwidth memory (HBM) and DRAM that store and rapidly supply the vast quantities of data that those processors work on.

Since the AI revolution started roughly three-and-a-half years ago, the scoreboard has been almost cartoonish. Since ChatGPT's public release, Nvidia stock has risen 1,290%, while Micron has soared 1,630%. These are not typical numbers, even for a bull market.

The reason behind their parabolic ascents is simple. Hyperscalers are spending more than $700 billion annually on AI infrastructure, and large slices of those checks are being allocated to GPUs, CPUs, HBM, and DRAM. Nvidia dominates the accelerator conversation, while Micron is one of only three companies that make the memory stacks that sit next to Nvidia's chips. As long as Amazon, Microsoft, Meta Platforms, and Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG) keep building data centers, Nvidia and Micron will continue cashing the invoices.

That story is not wrong, but it is incomplete. The quiet winner of the infrastructure cycle may be one of the companies writing a lot of these checks. Alphabet already dominates the consumer and enterprise demand side, the software side, and, increasingly, it's designing its own silicon. Wall Street has spent nearly four years treating all of that as a footnote. Warren Buffett and his new successor, Greg Abel, are not.

Image source: The Motley Fool.

Berkshire Hathaway is plowing into Alphabet stock

Berkshire Hathaway's first disclosed purchase of Alphabet stock showed up in the third quarter of 2025, a position of 17.8 million shares. The conglomerate sat on that stake through the end of the year before it accelerated its buying activity. During the first quarter of 2026, it nearly tripled its position in Alphabet. During the second quarter, Berkshire once again added shares on the open market and, more tellingly, wrote a $10 billion check for new shares as part of a larger private placement.

Buffett has been quite blunt about the origin story of Berkshire's position in Alphabet. "I initiated it," he told CNBC's Becky Quick during a recent interview.

Greg Abel, who succeeded Buffett as CEO at the start of 2026, has been aggressively adding to the position. Berkshire now holds around 106 million Alphabet shares worth nearly $38 billion. The position comprises about 13% of Berkshire's total stock portfolio, behind only Apple and American Express.

What makes Alphabet an attractive AI investment?

Nvidia sells the engine that powers AI models, and Micron sells the tanks for the fuel that keeps the engine running. Alphabet, by contrast, is building the car, the roads, and a growing share of alternative engines. Google Search is a money-printing machine, while YouTube draws the attention of billions of viewers. Meanwhile, Android sells smartphones and other consumer hardware around the globe. On top of these assets sit Gemini, Google Cloud, Tensor Processing Units (TPUs), and the Waymo autonomous vehicle fleet.

The recent performances of Google Cloud are something that should make chip bulls pay attention. In the first quarter, its revenue rose 63% year over year to $20 billion, with operating margins coming in at 33%. During the second quarter, sales from Google Cloud jumped 82% year over year to 24.8 billion, with an operating margin of 36%. Moreover, Google Cloud's backlog was a jaw-dropping $514 billion at the end of the second quarter. Cloud infrastructure is not a little side hustle for Alphabet -- it's turning into a second core profit engine alongside the core advertising segment.

On the silicon side of the cloud division are TPUs -- a type of custom silicon that can handle AI workloads at a lower cost than GPUs. Google designed them for its own specific AI workloads, and it used to keep the chips in-house, but now, it has started selling some to enterprise customers that want the custom silicon that trains Gemini models in their own data centers. This is another example of how Alphabet is turning what was once a cost center into a monetized product.

Waymo services now run on the order of 500,000 paid robotaxi rides a week across more than a dozen cities. A February funding round valued that business at $126 billion. As agentic AI applications enter wider production, Waymo could swiftly emerge as a business that contributes meaningful unit economics to Alphabet's broader ecosystem, much in the same way Google Cloud has scaled up over the last couple of years.

The thread stitching Alphabet's fabric together is Gemini. Search, YouTube, Android, Workspace, Cloud, and even the robotaxis increasingly run on a unified family of models. That structure is the difference between selling critical components and compounding an ecosystem.

Image source: Getty Images.

Alphabet stock looks like a bargain value

Alphabet's vertically integrated stack is how the company manages to remain consistently profitable. Cloud, which was actually losing money as recently as 2022, is now one of the juiciest parts of the company. The reason? AI has become the accelerant. Alphabet's companywide operating margin is around 34% and moving upward. That's impressive for a business that's spending like a utility building a new power grid.

GOOGL Revenue (TTM) data by YCharts.

The capital cycle is the objection everyone already knows. Alphabet has said it plans for between $195 billion and $205 billion in capital expenditures this year. While the company's free cash flow has gone temporarily negative and it has paused its stock buybacks, the bigger picture isn't as ugly as it might appear in a spreadsheet. In fact, this is simply a repeat of the same pattern Google Cloud already survived: Absorb operational pain, then watch the profit margin show up once the capacity you paid to build is fully subscribed.

On valuation, Alphabet stock is not priced like a company that is at the forefront of AI. The company trades at a forward price-to-earnings (P/E) ratio of 16. That is nearly identical to the long-run average forward P/E of the S&P 500. For a business with Search's moat, YouTube's scale, Cloud's backlog, its own accelerator roadmap, and an autonomous robotics play that has yet to contribute much more than a rounding error to its financials, this is not an expensive stock.

Don't get me wrong: Nvidia and Micron will keep winning every time a hyperscaler or neocloud orders another server. But smart investors realize they are not the only companies getting paid. Alphabet is a rare business selling the infrastructure, running frontier models, owning multiple distribution channels, and still generating the kind of profits Buffett spent a lifetime buying. Against this backdrop, I see Alphabet as a no-brainer opportunity to buy hand over fist and hold onto throughout the AI infrastructure era.

American Express is an advertising partner of Motley Fool Money. Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, American Express, Apple, Meta Platforms, Micron Technology, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

There is a reason the S&P 500 (SNPINDEX: ^GSPC), Dow Jones Industrial Average (DJINDICES: ^DJI), and Nasdaq Composite (NASDAQINDEX: ^IXIC) are among the greatest wealth creation vehicles in financial history.

With indexes, you don't need to time the market or pick the next Nvidia (NASDAQ: NVDA). You also don't need to hop from gold to oil to real estate every time an economic outlook shifts. Over decades, owning productive businesses has been one of the most reliable passive ways to build wealth.

Simply put, companies generate earnings, reinvest capital, create new products, authorize share buybacks, and issue dividends, all while growing alongside the economy. This compounding effect is difficult for alternative assets to replicate over a long-term horizon.

With that said, investing in the stock market is not always smooth sailing. The catch is that the stock market can be an incredible source of wealth while also becoming unusually expensive from time to time. Right now, the price investors are paying for the stock market is the key variable to note.

Image source: Getty Images.

The stock market is doing something for only the sixth time in 155 years

On the surface, investors have plenty of reasons to be bullish as the major indexes hover around record highs. Artificial intelligence (AI) has unleashed an unprecedented capital spending cycle around data centers, graphics processing units (GPUs), networking, memory, power, and everything else needed to build out the AI infrastructure stack. There is a potential glitch in the AI machine, however.

Some investors are becoming increasingly concerned that new Federal Reserve Chairman Kevin Warsh could hike interest rates. The math tells us that if rates move higher, the cost of capital rises as well. This matters for an economy that's pouring hundreds of billions of dollars into AI infrastructure on an annual basis. Since the AI build-out is one of the core pillars supporting the current market rally, anything that threatens the capex cycle's pace could fuel a nasty repricing.

To be sure, I'm not personally distracted by monetary policy decisions. Instead, I am laser-focused on valuation, and so it's natural to immediately look at the price-to-earnings (P/E) ratio to determine whether a stock is expensive or reasonably valued. The problem with this approach is that P/E multiples can be noisy.

The reason is that a company's earnings can fall during a recession or be boosted by a temporary boom. This means that looking at just one year of earnings can give an incomplete picture of what a business is actually capable of earning during the course of a full economic cycle. This is where the CAPE ratio comes in.

The cyclically adjusted price-to-earnings (CAPE) ratio, or the Shiller P/E, smooths out valuation noise by measuring stock prices with average inflation-adjusted earnings over a 10-year horizon. This makes valuation readings less distorted by one unusually good or bad year.

CAPE readings have been backdated to 1871, giving investors 155 years of historical context. The average CAPE level during this period is about 17.8. Currently, the CAPE's reading of 41.1 is more than double its long-term average.

S&P 500 Shiller CAPE Ratio data by YCharts.

What happens when the CAPE ratio rises?

There have been six periods when the CAPE ratio sustained a reading of 30 or more for consecutive months during a broader bull market. Right now is one of those times. Let's see how the previous five periods played out.

1. 1929: Stock prices peaked during the Roaring Twenties. For historians, what followed is obvious: the infamous 1929 crash and the devastating bear market that plagued the early 1930s.
2. 1997 to 2001: During this period, the CAPE reached its all-time high of about 44 as the dot-com boom accelerated. The tech-heavy Nasdaq soared for years, proving that a frothy market can become even more expensive. The optimism, however, was largely supported by nothing more than hope and euphoria. Eventually, the bubble burst, and the Nasdaq dropped roughly 77% from its peak to trough.
3. 2017 to 2018: The S&P 500 climbed for nearly a year before the market experienced a sharp sell-off during the fourth quarter of 2018.
4. 2019 to 2020: Stocks were soaring heading into 2020. But within the first two months of the year, the S&P 500 stumbled into a brief bear market as the global economy effectively shut down due to the COVID-19 pandemic. Although the recovery was extraordinarily fast, the timing and scale of the initial decline were severe.
5. 2020 to 2022: After the pandemic-driven crash, stock prices quickly recovered thanks to monetary stimulus and ultra-low interest rates. The rally eventually came to a halt when inflation peaked at about 9%, prompting the Fed to raise interest rates. The S&P 500 entered a bear market in 2022, and high-growth technology stocks in particular suffered deep losses.

What should investors do if the stock market crashes?

There is an important nuance in the analysis explored in this piece. Specifically, a CAPE ratio above 30 does not mean price will soon fall -- the market can remain expensive for a surprisingly long time. This is why valuation is best viewed as a warning light, not as a stop sign.

Although history suggests a disaster should follow an expensive market, it does not definitively tell us when such an event will occur. History also teaches investors that the stock market survives bubbles, recessions, wars, inflationary shocks, and financial crises -- always coming out the other side and reaching new highs.

^SPX data by YCharts.

This is the entire point of staying invested for the long haul. Investors can't control when a correction or crash will inevitably arrive. What they can control is how exposed they are when volatility comes. The biggest mistake an investor can make right now is to confuse an expensive market with a broken one. The same market that brings painful corrections also produces the recoveries that create long-term wealth.

It's a good idea to always keep some cash on hand instead of being 100% invested in stocks. By doing so, you can buy quality businesses at more attractive prices during a downturn. Trimming speculative positions while continuing to own high-quality businesses can help mitigate the blow should the market start to sell off. Diversifying to defensive plays outside the technology industry that has driven much of the market's recent rally is also a good way to hedge against dips in growth stocks.

Although a CAPE reading within shouting distance of all-time highs should make investors cautious about future returns, it shouldn't convince you that the stock market is headed for irreversible damage. Ultimately, the goal isn't to predict when a crash will happen. Rather, your goal should be to build a portfolio that can survive one, while anticipating when the next bull market inevitably begins.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Nvidia. The Motley Fool has a disclosure policy.

Wall Street has a habit of turning the occasional quirk into full-blown folklore, and few stories get recycled as hard as the September Effect. Every year around Labor Day, the same storyline makes its way into the headlines: Money managers come back from vacation, they start rebalancing their funds, and stocks subsequently take a hit.

For most of the market, this pattern brings unwanted selling pressure. For names that have already run hot for years, namely artificial intelligence (AI) chip stocks, it can feel like someone purposely yanked the rug from underneath things.

The question right now isn't whether the September Effect actually exists. It's whether they will deliver a lasting punch while the market leans hard on semiconductor leaders like Nvidia (NASDAQ: NVDA), Broadcom (NASDAQ: AVGO), Advanced Micro Devices (NASDAQ: AMD), and a handful of other silicon darlings.

Image source: Getty Images.

What does the September Effect actually look like?

Over the long haul, September is the only month that features a negative average return. Since 1928, the S&P 500 (SNPINDEX: ^GSPC) has lost 1.1% on average during the month of September and finished lower roughly 56% of the time. Meanwhile, the Nasdaq Composite (NASDAQINDEX: ^IXIC) has dropped 0.9% on average since its inception in 1971. What's interesting is that the Nasdaq has actually finished in the green 52% of the time, but the ugly years drag the index's long-term September average below zero.

The Dow Jones Industrial Average (DJINDICES: ^DJI) isn't any prettier. Since 1897, the Dow has dropped 1.1% on average during September and finished the month positive only 42% of the time.

As you can see, not every September is a bloodbath for stocks. Nevertheless, the pattern is consistent enough that traders treat this month as a legitimate seasonal headwind rather than a coincidence. The usual explanations are pretty straightforward: Portfolio managers rebalance after summer and get a head start on tax-loss harvesting. From there, it's just a self-fulfilling prophecy once enough investors decide to get defensive and rotate out of growth and into areas like consumer staples and utilities.

How have AI chip stocks performed during the September Effect?

The AI revolution includes a small number of September Effects so far: 2023, 2024, and 2025. To see how semiconductor stocks fared during these periods, I'll benchmark the category leaders against two popular chip-themed exchange-traded funds (ETFs).

The VanEck Semiconductor ETF (NASDAQ: SMH) is a concentrated basket of 26 chip names. Nvidia is the 800-pound gorilla, comprising roughly 23% of the fund. Other major holdings include Taiwan Semiconductor Manufacturing, Broadcom, Micron Technology, AMD, and ASML. The iShares Semiconductor ETF (NASDAQ: SOXX) is built around 30 AI chip stocks, including Nvidia, Micron, Intel, Marvell Technology, and Applied Materials.

In September 2023, both SMH and SOXX dropped about 7%. While that's pretty ugly, it beat Nvidia's decline of 10% and was on par with TSMC, AMD, and Marvell. In September 2024, both ETFs finished the month flat. While this underperformed the positive performances of major AI chip stocks, it was also insulated from the losses seen in select laggards.

Last year, the VanEck Semiconductor ETF gained 12% in September while the iShares Semiconductor ETF soared 11%. As you'd expect, the broader chip complex sported much higher gains.

Data by YCharts.

The lesson here is straightforward: Individual chip stocks can move much higher or much lower relative to a basket of stocks. These ETFs give investors exposure to the same AI chip theme, but come with less of the whiplash from single stocks.

Will the stock market correct this September?

The analysis in this piece is meant to drive home the point that September has never been a reliable signal that a correction or crash is on the way. At best, it is a seasonal phenomenon that sometimes shows up and drags the market down. The last two years proved that strong earnings, falling interest rates, and insatiable demand for AI compute can overpower the historical pattern.

Investors need to accept that AI chip stocks are growth names with frothy valuations and high liquidity. When the market gets uneasy, stocks that previously rallied hard are precisely the ones that get sold first. This is not the same thing as a fracture in the AI infrastructure thesis. Data center build-outs are accelerating, hyperscalers are investing in custom ASICs, and memory demand is not going away just because September arrived.

The practical move is not to dump all of your chip stocks this month. Instead, you need to decide how much volatility you can actually tolerate. If you own individual stocks, size them in such a way that a 10% drop doesn't force you to panic-sell.

If you want to participate in the AI chip theme without the risks that come with picking individual stocks, SMH and SOXX offer compelling upside while keeping downside pressure relatively insulated.

Remember, the market's worst month has a history of being followed by better ones -- I'm looking at you, Santa Claus rally! While the September Effect is real enough to respect, it should not be influential enough to dictate decisions that can affect your portfolio for the rest of the year.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends ASML, Advanced Micro Devices, Applied Materials, Broadcom, Intel, Marvell Technology, Micron Technology, Nvidia, Taiwan Semiconductor Manufacturing, and iShares Trust-iShares Semiconductor ETF. The Motley Fool has a disclosure policy.

By now, you probably don't need an elaborate explanation about Space Exploration Technologies' (NASDAQ: SPCX) initial public offering (IPO). The company's June debut was the largest IPO in history, raising roughly $85 billion at a valuation of around $2.1 trillion. For a brief moment, SpaceX was actually more valuable than Amazon. This part of the story is old news for those who have been following the stock, though.

What's more interesting is how SpaceX has traded since its IPO pop. Just days after the IPO, shares reached an intraday high of roughly $226. However, after the company's initial euphoric ascent, concerns about SpaceX's aggressive capital expenditure plans and the potential for post-IPO lockup expirations to pressure the stock fueled a flurry of selling prior to its first earnings report as a public company. After bottoming out at just under $105, shares have started to rebound again.

Still, SpaceX now trades roughly 34% below its post-IPO peak, and in the vicinity of the $150 per share price where it opened on its first day of trading. Investors may be wondering whether this is an opportunity to buy the dip or the precursor to yet another drawdown. While I don't have a crystal ball, I do have a useful data set that shows a consistent pattern among mega-hyped IPOs. Spoiler alert: The direction of SpaceX stock is anyone's guess. Investors who want to add it to their portfolios should buckle up and prepare for a bumpy ride.

Image source: Getty Images.

Analyzing blockbuster IPOs

The first IPO I am going to analyze is Palantir Technologies (NASDAQ: PLTR), which went public via a direct listing in September 2020. Shares opened at $10 and rocketed to a high of around $45 within the first year. At the time, Palantir was not seen as a darling of the artificial intelligence (AI) software complex. Instead, the company's early ascent was driven by meme stock era updrafts, fueled by Reddit users on the WallStreetBets forum. After that initial parabolic rise, Palantir spent most of 2021 giving back its gains and eventually settled in the mid-$20s range.

Snowflake (NYSE: SNOW) had a similar arc to Palantir, although the degree to which the stock moved was more dramatic and prolonged. The data warehouse specialist priced its IPO at $120, but shares actually opened their first day of trading at around $245. Before the end of 2020, Snowflake stock had surged to almost $400. While shares then sold off from this peak, Snowflake was still changing hands at prices of around $300 one year following its IPO.

Those two performances might suggest IPO investing usually leads to multibagger gains, but smart investors know there is more to this analysis. Figma (NYSE: FIG) is where the cautionary tale begins.

Figma stock opened at $85 last July and closed its debut session above $115. The very next day, shares spiked to nearly $143. Sounds great, right? Unfortunately for those who chased the momentum, Figma stock eventually crashed -- bottoming near $17 this spring. While Figma has started to show some signs of a comeback, the stock still experienced a peak-to-trough drawdown of more than 80%. That is absolutely brutal.

Cerebras (NASDAQ: CBRS) is a semiconductor company that went public earlier this year. While the stock hasn't reached its first anniversary as a public company, I still see the direction of its price action as useful information.

The Cerebras IPO was priced at $185, but early interest pushed its opening day first-trade price to $350. On that first day of trading, Cerebras stock touched $386. But over the last few months, shares have been all over the place -- ranging as low as $170 and swinging as high as $250. Its current price represents a drawdown of roughly 55% from its opening peak.

Tech IPOs tend to follow a similar path

Back in July, wealth management firm SCS Financial put together an interesting analysis featuring the performance of nearly two dozen IPOs across the technology and tech-enabled services landscapes.

The data includes offerings as far back as the late 1990s, when Amazon and Nvidia went public, as well as a number of early to mid-2000s names like Alphabet, Netflix, Facebook (now Meta Platforms), Tesla, and Uber. The most recent IPOs in the data set included, unsurprisingly, Snowflake, Palantir, and Cerebras.

The takeaway was that the stocks in this cohort experienced a median decline of about 53% from their post-IPO highs. The report also found that IPOs as a broader group have trailed the S&P 500 meaningfully over the last decade or so. For reference, since its inception in late 2013, the Renaissance IPO ETF has significantly underperformed the benchmark index. That ETF holds IPO stocks from recent years such as CoreWeave, Astera Labs, Reddit, and Rubrik. The fund holds onto its positions for up to three years before exchanging them for new IPO stocks.

IPO data by YCharts.

Where could SpaceX stock be trading by June 2027?

If I apply the same median 53% haircut to SpaceX's $226 peak, then a potential floor for the stock sits somewhere around $105 (which is about where it sat at its lowest point so far). On the more extreme end -- closer to what Figma experienced -- SpaceX stock could bottom closer to $45. I don't think that will happen, though.

If I put the entire peer group in this analysis together, a defensible range for SpaceX stock one year after its IPO could be something around $105 on the realistic bearish end, roughly $160 in a base case, and potentially upward of $200 in an upside scenario that mirrors those IPOs that have displayed the rare ability to reclaim and subsequently build on prior highs.

If you invest $10,000 today at SpaceX's current $147 share price, then it could be worth anywhere between the following by next June:

• Bear case ($105): Worth about $7,100 -- a loss of roughly 29%.
• Base case ($160): Worth about $10,900 -- a modest gain of roughly 9%.
• Bull case ($210): Worth about $14,300 -- a gain of roughly 43%.

While none of this is a perfect forecast, it does represent a series of plausible outcomes supported by comparably hyped IPOs. Ultimately, the analysis here serves as a reminder that even category-defining companies like SpaceX can be particularly risky short-term investments if you chase them when they're at the wrong altitude.

Adam Spatacco has positions in Alphabet, Amazon, Nvidia, Palantir Technologies, and Tesla. The Motley Fool has positions in and recommends Alphabet, Amazon, Figma, Meta Platforms, Netflix, Nvidia, Palantir Technologies, Reddit, Snowflake, and Tesla. The Motley Fool recommends Uber Technologies. The Motley Fool has a disclosure policy.

During Nvidia's (NASDAQ: NVDA) second-quarter earnings call, Jensen Huang used the word "visibility" only once. He did not spend an extended period of time arguing with skeptical analysts about the durability of the artificial intelligence (AI) build-out.

Instead, he explained that Nvidia can now see further upstream and downstream than it ever has -- into wafers, memory, optics, land, power, and the facilities that will house the next wave of AI systems. This single word, set against yet another quarter of record results and a supply-constrained outlook, is more meaningful than the bubble commentary that has followed Nvidia stock for over a year.

Jensen Huang: Image source: Nvidia.

Why does the AI bubble story exist?

The stance that the AI sector is in a bubble is nothing new. It is a story wrapped around circular financing deals, stretched balance sheets, and a lingering fear that demand is being manufactured by the same companies that are selling the picks and shovels. Hyperscalers and AI labs are spending enormous sums procuring accelerators that are then used to generate tokens. The resulting revenues from generative models and cloud infrastructure are subsequently used to justify the premise that more chips are needed. Critics see a loop that looks eerily similar to the fiber optic infrastructure build-out of the late 1990s, when installed capacity raced far ahead of profitable uses.

Skeptics also view the growing market for custom silicon to support the idea that Nvidia's moat is narrowing. If Amazon, Alphabet, Microsoft, and Meta Platforms design their own AI accelerators, then Nvidia's pricing power should erode, in theory. In addition, they point to free cash flow turning negative at some of the largest AI spenders as their capital expenditures surge ahead of their operating cash flows.

Against this backdrop, Nvidia starts to look like a fashion design that will fade after the first generation of data centers is fully depreciated and the second generation starts to look optional. At first glance, this logic may look sound. However, it is also incomplete. Bears are treating Nvidia as nothing more than a chip vendor waiting for purchase orders rather than as the company organizing the entire factory that produces artificial intelligence.

Nvidia just offered fiscal 2028 guidance early

Huang explained that by working with "land, power, and shell companies all around the world," Nvidia can better prepare "all of this computing that's going to be built that will ultimately deploy for our ecosystem and our customers."

That planning has given Nvidia unprecedented visibility, which is why the company was able to guide for 70% revenue growth for its fiscal 2028, which won't start until Jan. 31, 2027, even though it has historically refused to forecast that far ahead. Management made it clear that based strictly on the level of demand for Nvidia's products, it could deliver growth significantly greater than 70%. But given the supply constraints on the hardware that goes into its architectures, 70% growth is a floor the company believes it can deliver while keeping customers, shareholders, and the supply chain aligned.

Nvidia's visibility is not abstract in the slightest. The top five hyperscalers are expected to lay out nearly $800 billion on capital expenditures in 2026 and $1.3 trillion next year. Meanwhile, cloud backlogs are around $2 trillion. These figures matter because they are not being used for marketing. They are being published to support the case around build plans of the customers that account for half of Nvidia's data center business. The other half -- neoclouds, sovereign projects, industrial buyers, and enterprises, which are grouped as ACIE (AI clouds, industrial, and enterprise) -- is growing even faster and compounding at a pace that looks far different from a traditional fashion cycle.

When spending at this scale is tied to multiyear site development plans, power interconnects, and memory allocations, demand signals stop looking like quarterly swings and start looking like a city industrial plan. Nvidia's confidence to publish a forecast for its next fiscal year is the public company equivalent of that plan. Think about it: Bubbles usually don't form when manufacturers tell their entire ecosystem how much product they will actually be able to ship a year ahead of time.

Nvidia is moving from chips to factories

Smart investors are beginning to recognize how Nvidia is expanding beyond graphics processing units (GPUs) and central processing units (CPUs). The company is quietly building the architecture of AI factories: full-stack systems in which the CPUs, GPUs, networking, software, and the physical site are designed in unison so that each new product raises the revenue opportunity per gigawatt of power.

That opportunity is already on display as costs have grown from roughly $18 billion per gigawatt in the Hopper era to $25 billion with Blackwell and now $40 billion with Vera Rubin. Nvidia understands how incremental market share will come not from winning another server rack but from owning more of the entire factory.

This is where Marvell Technology, Nokia, and Coherent fit into the equation. Nvidia holds equity stakes in each of these companies, which bring custom silicon, radio-access networks (RAN), and optical interconnects onto one platform.

Marvell specializes in custom XPUs (specialized accelerator chips designed for specific use cases) and silicon photonics used in Nvidia's NVLink Fusion fabric. Nvidia's partnership with Nokia extends its reach into AI-RAN, turning edge devices into another platform for producing and consuming tokens. Meanwhile, Coherent supports the optical backbone that replaces copper connectivity products as chip clusters grow. Taken together, these relationships give Nvidia a line of sight into networking, photonics, and telecommunications demand that a pure-play GPU designer would never see.

This level of visibility changes planning all across the supply chain. Nvidia has already warned that rising memory prices will put pressure on its gross margins into next year. But because Nvidia sits so far upstream with the three major memory suppliers -- Micron Technology, SK Hynix, and Samsung -- and sees land, power, and shell demands years in advance, it can redesign architectures, lock in capacity and supply agreements, and set customer expectations before a shortage turns into a surprise.

The takeaway here is straightforward: AI spending can still be cyclical at the margin level as memory inflation will stress near-term profitability. But this is not the same thing as a bubble preparing to pop. A bubble bursts when demand evaporates because sentiment suddenly changes.

What Huang is describing is secular demand that is already booked across land, megawatts, and critical components, with Nvidia positioned to capture a rising share of each new factory rather than fighting to maintain its share of each incremental chip shipment. For investors with long-term time horizons, now looks like just as good a time as ever to scoop up some shares of Nvidia and hold onto them as the AI infrastructure era kicks into gear.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, Coherent, Marvell Technology, Meta Platforms, Micron Technology, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

Elon Musk has a history of stretching the horizon until it looks close enough to grab. In late August, he offered his "best guess" that Space Exploration Technologies (NASDAQ: SPCX) could generate roughly $3.5 trillion in revenue by 2033.

That prediction was posted casually on social media, not offered as formal company guidance. However, as is frequently the case when Musk communicates, his words still landed with force. This forecast invites some simple questions:

1. What does SpaceX actually sell today?
2. How big a leap is Musk trying to sell?
3. Why should investors treat a timeline this aggressive as anything more than marketing?

Let's dig into SpaceX's current state and explore Musk's history using timelines. Spoiler alert: History suggests he's nowhere near correct with this prediction.

Elon Musk. Image Source: The White House.

How big is SpaceX?

Despite its name, SpaceX is not primarily a launch business at this point. The company also has a connectivity segment built around Starlink's broadband service as well as an artificial intelligence (AI) infrastructure division.

Through the first six months of 2026, SpaceX generated $12.5 billion in total revenue -- an increase of 54% year over year. Starlink was the largest contributor, comprising 60% of sales. The space segment only generated about $1.6 billion in revenue during the first half of the year. Not only was that virtually flat year over year, but accelerating research and development costs for the Starship program have resulted in widening operating losses for the segment.

AI infrastructure encompasses the part of the company that leases capacity and related services to hyperscalers. This division consolidates xAI (the maker of Grok), X (formerly Twitter), and the software coding platform Cursor. Although AI is SpaceX's fastest-growing business, the company spent $23.6 billion in capital expenditures on that segment alone during the first half of the year. In other words, AI is absorbing enormous capital spending for only $3.4 billion of revenue.

Putting Musk's $3.5 trillion claim into perspective

Amazon and Walmart are the world's two largest companies as measured by annual revenue. The two of them together recorded about $1.47 trillion of sales over the last year. Musk's 2033 forecast is nearly 2.4 times that sum. It is also roughly 140 times SpaceX's current revenue run rate.

The consensus estimate of Wall Street analysts covering SpaceX is for total revenue of $44.6 billion for the company in 2026. If the company delivers as they expect this year, it would still need to grow at an 86% compound annual growth rate over the next seven years to reach Musk's $3.5 trillion projection. That aggressive outlook assumes no headwinds from regulations, emerging competition, or capital constraints.

Even the most blue-sky models from Wall Street analysts don't see the company reaching comparable revenue numbers until closer to 2040, and those predictions are tied to the following assumptions:

• Thousands of Starship launches per year.
• Starlink continues to add subscribers and government contracts at an accelerating pace, and enters the telecommunications industry -- possibly releasing its own mobile device.
• Orbital data centers become a real market.

Even if all these developments succeed on optimistic schedules, they don't automatically produce a company larger than today's two largest retail giants combined. Smart investors can see that Musk's forecast is not a modest stretch of SpaceX's current trajectory. Rather, it is a bold claim that SpaceX will converge upon several industries at once at a speed no other industrial company has ever sustained.

History is not on Musk's side

The reason I have major doubts about Musk's 2033 forecast is not just because space exploration and AI are hard businesses to compete in. It's also because Musk's track record on timelines at Tesla (NASDAQ: TSLA) has been one of overpromising and underdelivering for nearly a decade.

• 2016–2017: Full self-driving (FSD) was described as imminent, and a coast-to-coast autonomous drive was promised but never delivered.
• 2019–2025: Robotaxis were expected "next year," with 1 million robotaxis forecast for 2020. Neither happened. Unsupervised autonomous driving was repeatedly pushed to "next year" or "by year-end." Currently, Tesla operates an extremely limited, geofenced robotaxi service with safety monitors in Austin, Texas. This is far short of the original vision.
• November 2019-late 2021: Cybertruck was promised to be in production for $39,900 and feature a 500-plus mile range. Instead, volume deliveries began in late 2023 at roughly twice the expected price and for a vehicle with far less range.
• 2019-2022: The Tesla Semi was unveiled for 2019 production. A handful of pilot trucks didn't appear until late 2022.
• Recurring missed deadlines: A $25,000 Tesla car, early factory-scale Optimus robots, and solar-roof volume have all followed the same loop: an explicit date and years of slips, followed by another date.

Tesla's market value has long been supported by the story that these programs were just around the corner and would transform the company into something much larger than an automaker. In reality, the company's milestones have drastically lagged the narratives that Musk has spun.

I think the same marketing talent and a similar appetite for distant numbers now surround SpaceX. A $3.5 trillion sales figure by 2033 is less of a budget and more of a valuation hyperbole: Keep the horizon far enough away that every quarter of growth and every successful operational stress test can be interpreted as progress toward an almost unimaginable destination.

Musk's history at Tesla suggests that his proposed milestone target for SpaceX is nothing more than a fantasy. The prudent interpretation is that SpaceX is a respectable, fast-growing company aspiring to become a vertically integrated industrial empire spanning space, internet services, and AI development. Based on the evidence of Musk's own track record, however, SpaceX is a far cry from becoming a $3.5 trillion sales machine seven short years from now.

Adam Spatacco has positions in Amazon and Tesla. The Motley Fool has positions in and recommends Amazon, Tesla, and Walmart. The Motley Fool has a disclosure policy.

Nearly three decades ago, Nvidia (NASDAQ: NVDA) started off as a chip designer for enhancing graphics for video games. As it turned out, these chips were also unusually good at the kind of math that trains artificial intelligence (AI).

Over the years, Nvidia built accompanying software and systems that allow researchers and cloud hyperscalers to actually use these chips for more-advanced applications. The combination of fast-processing chips plus the tools to run them made the company the default supplier when large language models (LLM) took off a few years ago.

Its Hopper chips were the workhorses of the first AI wave. Management smartly reinvested the profits it made from Hopper into research and development. Subsequently, the company's Blackwell architecture hit the market and became another monster success.

The theme is that each generation of new chips made it cheaper and faster to train models and get inference deployments into production. Now, Vera Rubin is the next step in Nvidia's chip roster. Let's explore what makes it unique and why this product could be a game changer for the business.

Image source: Nvidia.

What does demand for Vera Rubin look like?

During the second-quarter earnings call, management guided for $108 billion in sales for next quarter. Chief Financial Officer Colette Kress said, "We see Vera Rubin accounting for about 20% of data center revenue in Q3." Considering that Nvidia's data center segment makes up more than 90% of the company's total revenue, it's reasonable to forecast Vera Rubin being on track for something close to $20 billion of sales in its first real quarter of shipments.

This is an unusually fast start. Management, which already has orders from every major hyperscaler, called Vera Rubin the fastest product ramp-up in its history. This matters because cloud infrastructure providers such as Amazon Web Services, Microsoft Azure, and Alphabet's Google Cloud -- as well as AI labs like OpenAI and Anthropic -- continue to pour unprecedented sums into data centers. The largest AI developers are expected to spend close to $800 billion on capital expenditures this year and $1.3 trillion next year.

To quantify what this translates to for the company, consider the following: Nvidia used to collect about $18 billion in revenue for every gigawatt of computing capacity it helped install with Hopper. With Blackwell, that figure rose to $25 billion. Kress says that with Vera Rubin, the company can reach $40 billion per gigawatt. The increase comes from selling more of the underlying AI rack -- accelerators, networking, and now its own processors -- rather than just the graphics chips.

How Vera Rubin changes the economics of AI factories

Nvidia is marketing the Vera Rubin system as one that delivers more useful work for each watt of electricity consumed. In turn, developers can meaningfully reduce the cost of generating each AI token compared with the prior generations of hardware. For more-sophisticated uses in agentic AI, these efficiencies are important.

What makes Vera Rubin unique is that it also includes a processor to sit beside the chip itself. This expands Nvidia's addressable market, because customers are no longer only buying chip clusters but rather designing a complete factory for producing intelligence alongside Nvidia.

As AI infrastructure keeps accelerating, the supplier that owns more of that factory should be positioned to capture a larger slice of every new data center. This is exactly why the order book for Vera Rubin is already so full and why Nvidia is already talking about 70% revenue growth for next year.

Is Nvidia stock still a buy?

The stock trades at a forward price-to-earnings ratio (P/E) of about 24. Nvidia itself described its fiscal 2028 sales outlook as limited by how many chips it can produce, not by how many customers want them. This is important to understand, because if supply improves even nominally, or if the new Vera Rubin systems sell more of the adjacent gear than anticipated, earnings could come in much higher than Wall Street is currently modeling.

NVDA PE Ratio (Forward) data by YCharts.

There are some risks when it comes to investing in Nvidia. The cost of memory is getting exponentially more expensive, which will pressure gross margins for a few quarters. Meanwhile, China remains an uncertain market.

Nevertheless, the combination of an estimated $20 billion contribution from a brand-new product in its first quarter, a rising take per data center watt, and management's admission that underlying demand is stronger than the 70% growth target suggests investors may be underestimating the company's future cash flow.

For long-term investors, this is the simple case: The AI infrastructure cycle looks far from finished, but Nvidia stock is priced as if it might be. For this reason, I see it as a no-brainer stock to buy and hold at its current price point.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

Every year, as Labor Day fades and portfolio managers return from vacation, the same conversation starts: Stocks tend to wobble in the ninth month, so maybe it's time to play a little defense. If you own shares of Nvidia (NASDAQ: NVDA), this discussion may hit a bit harder.

Nvidia has been the market's artificial intelligence (AI) engine for four years now. Since OpenAI launched ChatGPT to the public on Nov. 30, 2022, shares of Nvidia have climbed by 1,200%, and the company has become the most valuable business in the world by market cap.

NVDA data by YCharts.

Sounds great, right? Well, what this also means is Nvidia can be one of the first names institutional funds sell when they want to take gains off the table. While the calendar is not destiny, it is not background noise either.

Let's dig into what the September Effect actually is, how it lines up with Nvidia, and what smart investors can do instead of guessing over the next several weeks.

What is the September Effect?

The September Effect is financial jargon used to describe a historical pattern: Over the long run, stocks have posted weaker returns in September than in any other month. In fact, September is the only month that shows a negative average across the major indexes.

Since 1928, the S&P 500 (SNPINDEX: ^GSPC) has averaged declines of about 1.1% in September and finished the month lower than it started roughly 56% of the time. The Dow Jones Industrial Average (DJINDICES: ^DJI) shows a similar pattern over an even longer stretch, with an average September drop of around 1.1% and a winning month rate of only 42%. Lastly, the Nasdaq Composite (NASDAQINDEX: ^IXIC) has averaged a decline of about 0.9% in September over the last several decades.

Interestingly, the Nasdaq has actually finished higher 52% of the time in September since 1971. But when it does decline, those slides have been large enough to outweigh the slightly large quantity of wins and pull the long-run average below zero.

Understanding these patterns matters. The stock market is far from guaranteed to slump in the month of September. As the analysis shows, many Septembers end in the green.

Nevertheless, the month tends to attract outsize selling. Money managers come back from summer breaks and rebalance portfolios. Investor psychology also plays a role: After a few ugly Septembers, many people simply expect another one and act accordingly. When you layer on a Federal Reserve meeting, jobs data, and inflation prints, you get a month that feels much heavier than August.

The important thing to keep in mind here is perspective. A 1% average monthly loss does not constitute a bear market. It's a seasonal headwind. Treating the September Effect as a definite prophecy is one way that people wind up selling good stocks at the wrong time.

Image source: Getty Images.

What Nvidia investors need to know in September

If you've been paying attention to the artificial intelligence (AI) revolution, you know by now that Nvidia doesn't trade like a sleepy blue chip stock. Instead, the semiconductor giant is a high-expectation, high-volatility growth stock touching the biggest capital-spending infrastructure cycle in modern technology history.

Here is how Nvidia stock has fared over the last few Septembers:

• September 2023: (10.3%)
• September 2024: 1.7%
• September 2025: 7.1%

Nvidia investors have clearly experienced at least one nasty September in the recent past, but they have also enjoyed better Septembers as the AI story continues to accelerate. This dichotomy is my whole point. Nvidia's month-to-month performance is less about the calendar and more about whether investors are in the mood to own one of the most crowded, profitable names in the AI ecosystem.

When institutional funds want to lock in some gains, Nvidia is one of the easiest stocks for them to sell because it is liquid, has a large weight in indexes, and has already delivered the kind of generational run that makes profit-taking feel both responsible and inevitable.

Underneath these buying and selling dynamics, Nvidia's business is the real story: Demand for data center chips remains intense, profit margins have been robust, and the company has kept beating estimates that Wall Street once thought were impossible to meet. At the end of the day, some seasonal selling does not erode Nvidia's dominance. It just means the stock can become briefly cheaper for reasons unrelated to the volumes of GPU shipments.

There's one more wrinkle for 2026: This is a midterm year. Some of the market's best Septembers came during midterm election cycles.

With that said, there is some important nuance to examine here. According to Scott Rubner of Citadel Securities, midterm election years have amplified the September Effect in more recent history, with "the average path weakening through month-end before recovering in October and accelerating higher around Election Day into year-end." Ultimately, the upcoming elections don't necessarily make September safe, but history proves that the "September is always terrible" moniker is a fractured argument.

The warning Nvidia investors should be aware of

My warning to investors is not to confuse a seasonal average with a long-term trading plan. Sure, Nvidia stock can fall sharply in a month when the market is already nervous about interest rates, valuations, or concentration in a handful of megacap names.

If you are holding a massive position in Nvidia stock because you think it will never go down, September is a good time to admit that it does. Even a 5% drawdown would not be entirely shocking in a company this massive and widely owned.

If your thesis remains that Nvidia will continue to be a pick-and-shovel winner in AI infrastructure over the next several years, a weak September is nothing more than a weather report. It's far from a reason to abandon the stock.

Against this backdrop, long-term investors should stay the course and treat any dips as rare opportunities to add if their position size still makes sense relative to the rest of their portfolio. Alternatively, if Nvidia has become an outsize slice of your net worth, it may be wise to use weakness to rebalance rather than trying to time the bottom.

Selling now and buying back in October may look neat on a chart, but in real life, it means hoping you can make two perfect decisions in a row, paying extra taxes, and risking missing the bounce that almost inevitably will show up once the seasonal selling fades.

Ultimately, the best strategy is to exercise patience and have a firm plan. Remember why you own Nvidia stock to begin with, and figure out how much of a position you can own and still stomach watching it fall. Then let the mechanics of September play out.

The investors who make the most money over the long term are not the ones who try to outsmart the calendar. They are simply the ones who weathered the ugly months and kept themselves invested in quality businesses throughout.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Nvidia. The Motley Fool has a disclosure policy.

September has a reputation problem on Wall Street, and if you own a basket of artificial intelligence (AI) stocks, it is hard to ignore. Investors talk about the September Effect the same way people talk about the first week back at the office after a summer vacation: The pace picks up, patience is thin, and small problems suddenly feel bigger.

Since 1928, the S&P 500 (SNPINDEX: ^GSPC) has posted an average return of -1% in September. The Nasdaq Composite (NASDAQINDEX: ^IXIC) features a similar September decline since its inception in 1971.

The long-running theory behind September sell-offs is that as fund managers come back from vacation, they rebalance their portfolios -- taking gains in stocks that have rallied hard while harvesting losses from laggards.

The ironic part of the September Effect is that the month actually finishes higher almost as often as it finishes lower. It's just that long-term average returns get dragged down by the ugliest years.

For those invested in the AI complex, the question is not whether a calendar anomaly exists. It is whether the September Effect still applies once you zoom in on the specific stocks that have defined the AI revolution over the last few years.

Image source: Getty Images.

The September Effect vs. the AI boom

The AI trade did not really get going until late 2022, after OpenAI commercially launched ChatGPT. While small, this gives us a useful sample: the Septembers of 2023, 2024, and 2025. Interestingly, they do not tell a linear story.

September 2023

The S&P 500 dropped about 5%, while the Nasdaq fell closer to 6%. Nvidia, which had just entered the trillion-dollar club, lost more than 10%. Macroeconomic factors, including relatively high interest rates and the possibility of a government shutdown, would be enough to make most investors uneasy in a normal year.

But when you layer on the fact that no one had ever witnessed anything quite like the AI boom, investors simply decided they had seen enough rallying in tech stocks for the time being. It felt like the September Effect arrived right on schedule and rocked the most expensive names hardest.

September 2024 and 2025

September 2024 was modestly positive for the S&P 500 and the Nasdaq, with both indexes rising about 2%. Big tech also fared well, with the "Magnificent Seven" rising anywhere between 2% and 22% during the month.

The script changed entirely in September 2025. The Nasdaq Composite jumped nearly 6%, while the S&P 500 rose 3.5%. Once again, megacap tech collectively boasted a strong performance, with Tesla and Alphabet rising 33% and 14%, respectively. The seasonal weakness that everyone usually braces for simply did not show up.

^SPX data by YCharts

Indeed, three years is not a century of data. Still, recent history shows mixed results rather than a uniformly grim performance. When investors are already nervous, growth stocks often get hit hardest because they carry higher valuations and more concentrated ownership.

On the flip side, when sentiment is positive, these same stocks have proven to be perfectly capable of ignoring September storms. That is the key takeaway: The September Effect is nothing more than a tendency. It is not a law, and the largest AI stocks have already shown they can shrug off that tendency.

How should investors prepare for September?

Perhaps the most useful thing an AI investor can do this September is refuse to treat the month as an all-or-nothing event. What I mean by that is do not dump your whole portfolio simply because a statistic from 1928 makes you feel cautious. Moreover, it's not wise to sit on 100% exposure to stocks if a 5% sell-off would force you to dump something you actually like.

The more balanced approach is to maintain a modest cash buffer and use any weakness as an opportunity to buy the dip in companies that already have real earnings growth powered by the AI story. Some quality examples include Nvidia, Microsoft, Alphabet, and Apple -- each of which has spent the last three years proving it can generate enough cash flow to fund its infrastructure build-outs.

Those are the types of businesses you want to own more of if the market decides to throw a seasonal tantrum. By contrast, speculative names generally only work out when everything else is already going up. Those are positions you can afford to trim before the calendar even turns.

Don't be fooled: I am not suggesting investors should try to time the market. Instead, I'm simply acknowledging that September tends to create air pockets and that smart investors would rather be buyers than forced sellers. The last three Septembers prove that these air pockets are not guaranteed in the AI era. But either way, having some dry powder and a short list of high-quality blue chip names you would happily buy at lower prices can keep September from turning into a point of regret.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, Nvidia, and Tesla. The Motley Fool has positions in and recommends Alphabet, Amazon, Apple, Meta Platforms, Microsoft, Nvidia, and Tesla. The Motley Fool has a disclosure policy.

Earnings season is just about wrapped up, and for the artificial intelligence (AI) crowd, the script was unusually consistent. Four of the most powerful companies in the technology world spent meaningful airtime during their earnings calls talking about one input to the cost structure: memory.

Not models. Not load power. Not even graphics processing units (GPUs), at least not at first. High-bandwidth memory (HBM), DRAM, and NAND were in the spotlight during various earnings calls. In a way, this is ironic, as these products used to be viewed as common commodities. Now, CEOs are treating memory and storage like scarce precious metals.

The story is straightforward: The companies writing the biggest checks just told investors that memory prices are still going up, supply remains tight, and they are going to keep spending anyway. For investors who own shares of Micron Technology (NASDAQ: MU) or Sandisk (NASDAQ: SNDK), this should be music to your ears.

Image source: Getty Images.

Four CEOs, one common discussion

Tim Cook didn't hedge during his final earnings call as Apple's (NASDAQ: AAPL) CEO. He told investors that Apple paid more for memory during the March quarter than in the December quarter, and subsequently expected to pay "significantly more" in the June quarter. He went on to say that he expects to pay an even higher premium in the September quarter, describing the memory supercycle as "a 100-year flood." Apple raised Mac and iPad prices to partially offset rising memory costs. While that helped offset some of the toll rising memory costs took on its margins, Cook made it clear that pricing dynamics could hit Apple even harder in the current quarter.

Elon Musk was more blunt. During Tesla's (NASDAQ: TSLA) earnings call, he singled out Micron by name, thanking the chipmaker for giving Tesla a "very significant allocation on reasonable terms given the pretty insane pricing of memory these days." During the SpaceX earnings call, Musk did his best to quantify what the memory landscape currently looks like. He suggested that memory production is rising by around 20% a year, which would be a great accomplishment under normal circumstances. However, if that level of growth is occurring while demand is rising by, say, 200%, then Economics 101 makes it clear that prices must rise.

During Amazon's (NASDAQ: AMZN) second-quarter earnings call, investors learned the company had increased its capital expenditure budget for the year from about $200 billion to $220 billion. CEO Andy Jassy explained that "the higher cost of memory" had pushed its capex requirements higher, but even with that extra $20 billion, the hyperscaler still won't be able to build enough compute capacity this year to meet its existing demand. Jassy expects that compute shortage to persist into 2027.

Jensen Huang's version of the memory story was on display during Nvidia's (NASDAQ: NVDA) fiscal 2027 second-quarter earnings call last week. The giant chipmaker sees demand for its processors that would support far more than 70% revenue growth in its fiscal 2028. But current supply dynamics support the 70% forecast. To ensure Nvidia can keep up with demand, the company has locked in $279 billion of supply and capacity commitments through its fiscal 2032, "primarily related to the procurement of memory." Of that figure, $267 billion relates to supply contracts for the next two and a half years.

The theme here is that four different companies explicitly described the same bottleneck as it pertains to their artificial intelligence (AI) build-outs: constrained supplies of memory.

The AI capex cycle isn't drying up; it's just getting more expensive

Here's the part of the AI story some investors keep getting backward. Rising memory prices are not proof that the AI infrastructure build-out is about to slow down. Rather, it's evidence that even the largest buyers still can't procure enough silicon even after memory prices have already exploded. If this were a fading cycle, big tech players would be cutting their spending. Instead, they are doing the opposite.

Amazon raised its capex budget purely because memory costs more now and because the company still needs more racks. Nvidia is guiding for unprecedented growth, yet still says the real constraint in the chip value chain is memory supply, not GPU demand. Apple is paying a premium for HBM and DRAM, and simply passing some of its higher costs on to customers. Meanwhile, Tesla and SpaceX are grateful just to get invitations to the party.

To me, this is what a durable supercycle looks like: Capex remains elevated because the opportunity cost of frugality would be to leave demand on the table. Memory supply tightness is not a mere side effect sitting alongside the AI boom. Memory is the boom, and it's showing up in big tech's receipts.

Why is this good news for Micron and Sandisk?

For now, the pain that the tech sector is feeling over soaring memory prices is real, and it's showing up in two obvious places: gross margins and free cash flows.

Apple has already experienced a sequential drop in its gross margin due to soaring memory costs, while Nvidia is doing damage control ahead of time -- walking investors down from a gross margin profile in the mid-70s percentages toward the low-70s percentages as memory inflation works through its supply chain.

Meanwhile, Amazon and Tesla have experienced weaker cash conversion as they accelerate their capital expenditures and build more data centers or buy more chips. Many hyperscalers' free-cash-flow figures got ugly during the same quarter they were bragging about demand for their compute. That is the opportunity cost of building this infrastructure. Companies are laying out more cash up front, and the chips inside their servers cost more than their models assumed even just six months ago.

Someone is on the other side of this trade. Micron is selling into a market where only three DRAM suppliers have real pricing power, and HBM is sold out in advance going into next year. Sandisk specializes in NAND flash, prices of which are getting dragged higher as capacity budgets move downstream toward storage.

MU PE Ratio (Forward) data by YCharts.

The takeaway here is that the AI capex cycle is very much intact and squeezing the memory buyers. Higher memory prices and bigger data center build-outs are two sides of the same invoice. The companies paying the invoice are already telegraphing a warning. The companies printing the invoice -- Micron and Sandisk -- are the ones the warning is for. With both of these memory stocks trading at modest forward price-to-earnings (P/E) multiples between 6 and 7, it could be that Micron and Sandisk remain grossly undervalued. In that context, they represent compelling opportunities to buy now and hold as the AI infrastructure era matures.

Adam Spatacco has positions in Amazon, Nvidia, and Tesla. The Motley Fool has positions in and recommends Amazon, Apple, Micron Technology, Nvidia, and Tesla. The Motley Fool has a disclosure policy.

Nvidia (NASDAQ: NVDA) sits at the center of two very different bets on artificial intelligence (AI) infrastructure. One side of the equation is loud and already consumes Wall Street's attention: a deepening commercial and financial partnership with OpenAI. The other side is quieter and much easier to miss: a reported agreement to acquire Hugging Face, a website where open-source AI models are published and shared.

The Nvidia-OpenAI relationship builds and fills AI factories. With Hugging Face, Nvidia can own something far more lucrative -- the marketplace where AI factories get their orders. Let's break down both deals and explore what's at stake for Nvidia investors.

Image source: Nvidia.

Nvidia's relationship with OpenAI bridges compute and credit

In February, OpenAI conducted a funding round that raised $110 billion at a $730 billion pre-money valuation. During that round, SoftBank and Nvidia each invested $30 billion, while Amazon committed $50 billion. More recently, SoftBank's SB Energy announced that it plans to build a data center at PORTS-Pike Technology Campus in Pike County, Ohio, and lease the facility to OpenAI.

The initial build-out plan covers 4.25 gigawatts of capacity, with an option for an additional 3.75 gigawatts. Nvidia will be the exclusive compute supplier as capacity comes online in 2028. According to an 8-K filing in late August, Nvidia entered into a series of residual value guarantees for up to $105 billion to help finance this infrastructure.

On the surface, this deal might look like a home run for Nvidia. But think about this for a minute: OpenAI needs graphics processing units (GPUs) from Nvidia to train its models. At the same time, OpenAI is hemorrhaging cash. So now, Nvidia is stepping in as a financier to bankroll OpenAI's infrastructure roadmap so that OpenAI can, in turn, buy more chips from none other than Nvidia. Unsurprisingly, Wall Street is skeptical of the mechanics, with some calling this financing arrangement too circular.

Nvidia CEO Jensen Huang and CFO Colette Kress both addressed this concern during the recent fiscal 2027 second-quarter earnings call. Huang described the arrangement with OpenAI as a way to lock in demand for Nvidia compute so OpenAI can build productive AI factories that can be "upgraded repeatedly." Kress went even further, saying that the downside risk from the OpenAI financing is limited because demand from frontier labs is so strong that they create entirely new businesses and ecosystems for Nvidia. She stated that "equity returns on our invested capital will be excellent."

Nevertheless, Wall Street still sees a supplier helping one of its core customers finance the very buildings that will be packed with that same supplier's hardware -- a loop that undoubtedly makes credit desks skittish.

A transformative acquisition smart investors won't overlook

While the arguments around Nvidia's tie-up with OpenAI echo, recent reports suggest that Nvidia is acquiring Hugging Face for $12.9 billion. Despite its funny name, Hugging Face is not some sort of mysterious AI start-up. It's really just a public library and developer workshop unified under one roof.

Developers who train AI models don't keep them hidden on private laptops. Instead, they can upload them to Hugging Face in much the same way a programmer can upload code to GitHub. This makes it easier for other developers to find, download, improve, and publish their own versions of the same model -- making Hugging Face a natural resource for sharing AI development work instead of starting from scratch.

Why is the Hugging Face deal more important than the OpenAI partnership?

With OpenAI, Nvidia gets a single enormous customer and years of exclusive capacity at one data center location. It also increases the chipmaker's customer concentration risk. OpenAI and many of Nvidia's largest customers are already exploring custom silicon designs. A guarantee on one campus in Ohio does not mitigate that risk entirely. It only ensures that those particular buildings will be full of Nvidia's hardware.

Hugging Face is a completely different kind of asset that strengthens Nvidia's entire ecosystem. CUDA is Nvidia's software layer, and includes an extensive library of tools that let programs use the full parallel-processing power of the company's GPUs. Nvidia's tight integration between its chip architectures and CUDA creates a legitimate lock-in with developers. This is more valuable than any single contract.

If Nvidia owns the domain where open-source models are most commonly posted, discovered, and turned into productive applications, it gains visibility into which models are actually winning. From there, Nvidia can fine-tune its hardware and software for the workloads that are shipping. While closed-system labs can try to switch to competing platforms, the open-source world is far more fragmented and already largely lives on Nvidia's architecture. Owning the sharing layer is how Nvidia can keep its competitive moat.

The OpenAI partnership comes with an enormous order for chips that have not been installed yet because the facility they're headed for hasn't been built. In contrast, acquiring Hugging Face provides Nvidia a faster path to own the marketplace that will determine which chips get ordered as new frontier labs join the current generation's leading platforms. In my view, this is why the Hugging Face deal provides a longer-runway opportunity for Nvidia.

Adam Spatacco has positions in Amazon and Nvidia. The Motley Fool has positions in and recommends Amazon and Nvidia. The Motley Fool has a disclosure policy.

Calendar dates are not supposed to matter in efficient markets. Stock prices should reflect a company's cash flows, as well as sentiment toward macroeconomic variables such as interest rates. The fact that Labor Day is right around the corner shouldn't matter. Yet for nearly a century, one month has exhibited a noticeably different pattern from the rest of the year.

September is the only month in which the S&P 500 (SNPINDEX: ^GSPC) has posted a negative long-run average return, down roughly 1.1% between 1928 and 2025. While there is no single reason for September's historical weakness, much of the declines are attributed to portfolio managers returning from summer vacations and rebalancing their portfolios. This means liquidity that had thinned out during July and August has to absorb new selling pressure.

Investors should care because the September Effect is influential enough to manifest in even the most crowded growth trades. Let's explore what that could mean for artificial intelligence (AI) stocks.

Image source: Getty Images.

What is the September Effect, and why does it matter?

One practical reason to notice patterns around September is concentration risk. When a handful of richly valued growth stocks dominate the S&P 500, a seasonally weak month can turn into a harsher de-risking event rather than a gentle period of digestion. That is the setting in which AI stocks now sit. More specifically, the largest AI companies, known as the "Magnificent Seven" -- Nvidia, Apple, Alphabet, Microsoft, Amazon, Tesla, and Meta Platforms -- now make up nearly 34% of the S&P 500's value.

Since OpenAI commercially launched ChatGPT and ignited the firestorm in AI stocks in November 2022, three Septembers have passed. Spoiler alert: They did not tell a unified story.

^SPX data by YCharts

In September 2023, both the S&P 500 and Nasdaq-100 indexes fell roughly 5%. While the performance of the Magnificent Seven was widely distributed, Nvidia, which is the clearest proxy for the AI revolution, dropped the most at 10%. The following two Septembers looked quite different. In 2024, the S&P 500 and Nasdaq-100 each rose around 2%. Meanwhile, in 2025, the S&P 500 gained 3.5% while the Nasdaq-100 advanced 5.4%.

While the September Effect did not vanish, it hasn't necessarily been a reliable wrecking ball for the AI boom. This mixed record over the last few years is an important detail to understand. AI companies are still growth stocks with high duration. This means these stocks can fall further and faster than average, even if the overall September anomaly is modest. This vulnerability was on display in 2023.

However, both 2024 and 2025 demonstrated that a supportive interest-rate environment, continued capital expenditures (capex), and compounding earnings can outweigh seasonality. While the last three years are not a large sample size, it can still serve as a reminder that the AI trade is not immune to the same September pressures that have long hit high-beta growth stocks.

Where should investors look beyond tech stocks in the month of September?

The best places to store capital during weaker periods are not mysterious. When investors rotate out of growth, they usually seek out mundane assets such as gold, short-duration Treasuries, and traditional defensive stock groups -- utilities, consumer staples, and parts of healthcare. According to research from Dorsey Wright, gold generated positive returns 58% of the time in September since 1987.

The reason is simple: These asset classes do not need a narrative about the next generative model release and tend to be less vulnerable when narratives around geopolitics or monetary policy change. When institutional investors trim winners and harvest losses, they tend to sell the assets that have rallied the most and are most liquid. That description perfectly fits megacap AI stocks after a multiyear run. Defensive cash-flow stocks, gold, and high-quality bonds are where capital gets deployed when the same sellers want to stay invested without adding unnecessary risk.

The investment takeaway here is pretty unglamorous. I would not liquidate a long-term AI or technology allocation simply because the calendar turned to September. Instead, I'd treat the month as a stress test of position size. In other words, if a small handful of semiconductor and platform stocks now dominate your portfolio in the way they dominate the S&P 500, a modest trim into momentum, paired with a balance in safe-haven assets, is a way to respect the historical odds without pretending to time the market.

At the end of the day, rebalancing your portfolio after a strong August is precisely the behavior that creates the September Effect in the first place. Doing it deliberately, rather than having it done to you, can make all the difference.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, Nvidia, and Tesla. The Motley Fool has positions in and recommends Alphabet, Amazon, Apple, Meta Platforms, Microsoft, Nvidia, and Tesla. The Motley Fool has a disclosure policy.

Over the last year, artificial intelligence (AI) memory stocks have been on fire. Expanding model sizes and more sophisticated agentic AI use cases are fueling a parallel boom between graphics processing units (GPUs) and the high-bandwidth memory (HBM) layered on top of these accelerators.

With its shares up by 697% over the last year, Micron Technology (NASDAQ: MU) has been one of the clearest beneficiaries of this memory boom. While the stock's parabolic ascent has been tough to ignore, some investors can't help but wonder when the memory trade will fade. After all, memory has historically been a highly cyclical market.

Well, Nvidia (NASDAQ: NVDA) may have just quietly signaled a tell about when Micron's business -- and its stock -- may peak and then begin to decline.

Image source: Micron Technology.

How to interpret Nvidia's supply and capacity commitments

In conjunction with its second-quarter earnings report, Nvidia's CFO commentary included a table outlining future supply and capacity commitments. "Our commitments increased from $119 billion last quarter to $279 billion, primarily related to the procurement of memory," she wrote.

Per Nvidia's forecast, the company has committed to lay out money for memory and other supplies along the following time frame:

• Remainder of fiscal 2027: $92 billion.
• Fiscal 2028: $87 billion.
• Fiscal 2029: $88 billion.
• Fiscal 2030: $6 billion.
• Fiscal 2031: $5 billion.
• Fiscal 2032+: $1 billion.

Investors can see that Nvidia is budgeting $267 billion on memory through the company's fiscal 2029 (which ends January 2030). Let's make one thing clear: Nvidia's purchase orders do not all translate into revenue for Micron. First, some of these funds will go toward other types of components and supplies. But in the tight memory market specifically, Nvidia works closely with the other two major players in the space, SK Hynix and Samsung, both of which are leading producers of HBM and DRAM.

What Nvidia's commitments schedule underscores is that memory suppliers now demand multiyear visibility and commitments from large buyers. For now, Nvidia has given them roughly three years of it. Fiscal 2030 is where Nvidia's leverage as a customer could become more evident. What I mean by that is Nvidia may not need to lock in HBM purchases for 2030 at today's premium prices; hence, the company has not yet outlined meaningful spending deals beyond fiscal 2029.

Understanding fiscal year timelines

Nvidia's fiscal years end in January, while Micron's end in August. Nvidia's last certain year of meaningful memory spend will happen between February 2028 and January 2029. This period straddles the back of Micron's fiscal 2028 and the early part of its fiscal 2029.

One important thing to understand is that shipments lag commitments. This means Micron can still package products in the middle of 2029 against orders Nvidia already placed. With all of that said, it's fair to say Nvidia will still need memory beyond 2029. A lean order book is not concrete evidence that memory demand is destined to fall off a cliff by 2030.

Processors bought during the current data center build-out will eventually need upgrades and be replaced by new architectures. In turn, these AI accelerators will continue to require HBM stacks. That level of demand is what's not showing up in Nvidia's supply and capacity commitments right now.

This is the trap to be aware of: Micron's revenue can -- and probably will -- look fine while its underlying demand trends are potentially headed for a slowdown. Starting in February 2029, the memory story may hinge more on negotiations over volume and price than it does today. If HBM is still in short supply relative to demand, then Micron will have the negotiating power. If not, then Nvidia will be able to spend less even as it keeps designing new, more powerful processors with greater memory demand.

When could Micron stock begin to see some pressure?

Remember, markets are forward-looking. Investors are not going to wait until 2030 to assess whether the supply-and-demand dynamics have actually changed. With that said, markets also do not look three years ahead while earnings are still compounding, like they are for Micron. This is why the stock could continue to rise throughout 2027 and 2028 even though Nvidia's commitment table is already public knowledge. As of now, the memory supercycle still has years of contracted shipments locked in.

February 2029 is the moment Nvidia's low-commitment year moves inside the window where the market could actually change how it prices Micron. If Nvidia does in fact scale back its purchase orders, investors will stop applying a scarcity premium to the AI memory market. My prediction is that Micron stock could peak somewhere between December 2028 and January 2029 unless Nvidia ratchets up its commitments well ahead of fiscal 2030.

Ultimately, Nvidia's memory commitments -- and the lack of them after 2029 -- do not tell us for sure that the memory market will dry up by 2030. But unless Nvidia makes it explicit that it will need more HBM going into the next decade, and that it's paying shortage prices for that HBM, I'd expect a de-rating in Micron stock. More specifically, smart investors could start trimming their positions between December 2028 and January 2029, leaving the rest of the market holding the bag and fueling a sharper drawdown thereafter.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Micron Technology and Nvidia. The Motley Fool has a disclosure policy.

One aspect of market psychology is that investors often think in terms of one quarter at a time. Nvidia (NASDAQ: NVDA) may have broken this habit in its fiscal 2027 second-quarter report. During the earnings call, management provided something public companies rarely give quite this early: a forecast for next year's growth.

Below, I'll detail why Nvidia's growth is redefining the debate about whether the artificial intelligence (AI) build-out is late-cycle theater or still in the early innings. Moreover, the analysis will touch on two important points that have haunted Nvidia stock for nearly a year: how much of this growth forecast does Wall Street actually believe, and whether Nvidia needs to regain its position in China's market to keep its empire running.

Image source: The Motley Fool.

What was Wall Street expecting for Nvidia's fiscal 2028?

Nvidia's preliminary fiscal 2028 outlook is straightforward. Revenue is expected to rise 70% year over year. Chief Financial Officer Colette Kress framed that figure as "supply constrained." Chief Executive Officer Jensen Huang made it clear that demand for the company's processors is growing by more than that. In essence, a 70% growth rate is what Nvidia's supply chain can "confidently deliver," especially with shortages of memory and other parts of the AI chip stack creating bottlenecks to production.

For reference, Wall Street analysts were modeling about 44% revenue growth for Nvidia's fiscal 2028, which begins Jan. 31, 2027. Given the Street's consensus estimate of $397 billion of revenue in fiscal 2027, that implies fiscal 2028 sales of roughly $574 billion. When applying Nvidia's 70% forecast rate to the same starting base, next year's expected revenue sits closer to $675 billion.

Here's where it really gets lucrative: If I use a higher fiscal 2027 revenue figure based on Nvidia's current run rate, the implied sales for next year land closer to $700 billion. In either case, the gap between Wall Street's expectations and Nvidia's new reality is roughly $100 billion in revenue.

What's astounding is that Nvidia is no longer growing off a small base of data center sales. Given its current trajectory, Wall Street must accept that a business already measured in hundreds of billions of dollars of annual sales can go on to add yet another several hundred billion in growth in just a single year. Under these conditions, procuring GPUs is no longer the constraint for AI training and inference. Instead, the pain points revolve around high bandwidth memory, packaging, power supply, and land.

Does Nvidia need China?

During the fiscal second quarter, Hopper-architecture products shipped to China accounted for less than 1% of Nvidia's data center revenue. Moreover, these shipments were dilutive to Nvidia's gross margin. Kress made it clear that "given ongoing geopolitical uncertainty, there is no China data center compute revenue in our forward outlook."

Think about that for a minute: Nvidia expects to generate $108 billion in sales in the third quarter, alongside a 70% growth rate in 2028, and the Chinese market is merely an option rather than a core pillar supporting the company's sales foundation. This matters for a few reasons. First, this level of growth, excluding China, discredits a convenient bearish argument that Nvidia would need a large presence in that market to sustain its dominance in AI processors.

Second, and more subtly, Nvidia's growth outlook over the next 18 months underscores that AI labs, neoclouds, enterprises, and sovereign buyers are becoming just as important as the hyperscalers. Kress quantified the non-hyperscale cohort as representing "roughly half of our data center business." When demand for its wares is this broad, the fact that it continues to cede ground in an important market like China is only a disappointment, not a thesis-killer.

Third, Nvidia's position in data centers remains undeniable, despite increasing competition from Advanced Micro Devices and custom silicon designers like Broadcom. Nvidia's outlook suggests the company is still fighting effectively to win incremental server demand in a contested AI infrastructure landscape. A vendor in Nvidia's position does not "need" China the way a competitor like AMD needs to prove it can expand globally at a comparable scale. For Nvidia, China is purely a source of incremental dollars and a strategic hedge, not a key engine powering its future growth.

Is Nvidia stock a good buy?

Nvidia stock trades at a forward price-to-earnings (P/E) ratio of about 23. This is a rather modest valuation compared to the highs it reached during the early cycles of the AI revolution. When paired with the company's reaccelerating data center growth, it's hard not to see Nvidia as a terrific value right now.

NVDA PE Ratio (Forward) data by YCharts.

But take a look at Nvidia's price/earnings-to-growth ratio (PEG ratio) as well. The PEG ratio measures a company's price relative to its expected future earnings growth. As a rule of thumb, any positive PEG ratio of less than 1 suggests a stock is undervalued. Currently, Nvidia's PEG is around 0.6. To me, it's clear the market is not paying up for the earnings path that Nvidia's management just outlined.

The takeaway here is simple: Nvidia's multiyear guidance is not a victory lap. It's a declaration that the bottlenecks to the AI build-out revolve around physical components, and that a meaningful return to the Chinese market is not something that the company would need in order to achieve a financial performance that the Street is under-predicting by a mile. At a forward earnings multiple that has somehow compressed even as the company's earnings power continues to compound, Nvidia stock is still worth owning.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Advanced Micro Devices, Broadcom, and Nvidia. The Motley Fool has a disclosure policy.

The September Effect is a persistent calendar anomaly in the capital markets, as stocks tend to deliver weaker returns in September than in any other month. This matters because it is not merely a statistical oddity. Institutional investors, mutual funds, and individual traders treat September as a period of heightened risk, a perception that becomes self-reinforcing.

After the quiet summer trading in July and August, portfolio managers return from vacation, reassess positions, and often trim winners or dump laggards ahead of the fourth quarter. Some funds operate on fiscal years that end in September or October, creating a short window for tax-loss harvesting.

The combination of lower liquidity, renewed scrutiny of valuations, and a psychological shift from summer complacency to autumn caution historically produces an average decline in the Nasdaq (NASDAQINDEX: ^IXIC), even though September often finishes higher.

Let's explore how the Nasdaq has held up in September and assess what it could mean for two darlings fueling the artificial intelligence (AI) revolution: Nvidia (NASDAQ: NVDA) and Micron Technology (NASDAQ: MU).

Image source: The Motley Fool.

How does the Nasdaq hold up during September?

Since its creation in 1971, the Nasdaq has averaged a return of -1% in September. This figure stands out because it is the only month with a negative long-term average in the index. Interestingly, September is not a consistent loser based on frequency, though. The Nasdaq has finished September higher roughly 52% of the time.

The apparent contradiction is clearer in the magnitude of the Nasdaq's September declines. When the index rises in September, the gains tend to be modest. But when it falls, the declines are often sharper and more concentrated -- dragging the overall long-term average down.

The result is a month that feels more risky than the actual percentage drop suggests. In turn, volatility clusters, liquidity thins, and ironically, the same names that fueled prior rallies often become the easiest sources of cash when portfolio managers decide to reduce exposure.

How seasonal weakness could impact AI stocks

Portfolio managers who have enjoyed outsize gains through the summer may use September's historically weak backdrop as an opportunity to lock in profits. Because mega-cap AI stocks are so liquid, they become convenient vehicles for selling. A modest rotation away from names like Nvidia and Micron can reduce the Nasdaq's upward momentum, which in turn can snowball into even more selling from retail investors.

Smart investors see how the same characteristics that generated sharp advances in the first place can also produce faster, harsher drawdowns at the flip of a switch. In a month already prone to larger-than-average losses, concentrated exposure amplifies the September Effect rather than cushioning it. The risk is not that Nvidia and Micron suddenly lose their long-term thesis. Rather, it is that short-term positioning and seasonal caution can override sound fundamentals during this period.

Nvidia and Micron bookend the month of September

In my eyes, Nvidia is the clearest proxy for the AI infrastructure supercycle. The company's data center GPUs power the training and inference clusters that hyperscalers like Alphabet, Amazon, and Microsoft, as well as a rising number of private enterprises and sovereign governments, are racing to build.

Nvidia reported its fiscal second-quarter earnings results in late August -- delivering another record quarter of explosive revenue and profit growth and raising the bar for the rest of the year and next year, too.

Nvidia's earnings arrived just days before September began, leaving investors to digest both the reported numbers and the seasonal calendar simultaneously. While a strong earnings report can support a stock price, it can also invite profit-taking in this particular instance, given the timing.

Micron sits in a different but equally critical position. Over the last year, high bandwidth memory (HBM) has emerged as a binding constraint on AI system deployments. Micron is one of only three companies with the scale and process expertise to supply the memory bottleneck. The company is scheduled to report fiscal fourth-quarter earnings on Sept. 30.

This timing places Micron's quarterly update after several weeks of potential seasonal selling pressure and after Nvidia's own earnings report has already set the new tone for the AI complex. If portfolio managers have started trimming AI-related holdings, Micron's results will be set against a more skeptical backdrop. Conversely, any confirmation of tight memory supply and rising pricing power could stabilize the AI trade just as September closes.

The prudent takeaway is not to sell blindly simply because the calendar changed. A smart approach is to size your positions so that a sharper, faster drawdown in high-beta stocks remains tolerable. Moreover, investors must be able to distinguish between post-earnings digestion and portfolio rebalancing, and a genuine change in the AI infrastructure or memory thesis. Remember, seasonality is just context, not a guaranteed signal to abandon a multi-year build-out story.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, Micron Technology, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

Nvidia's (NASDAQ: NVDA) second-quarter 2027 (ended July 26, 2026) earnings report made the usual point in unusually large numbers. Data center revenue reached $89 billion during the quarter, up 18% from the prior quarter and 117% year over year. What I think matters more than the headline figure is the bifurcation inside that number.

Hyperscale customers generated $48.7 billion in revenue, up 13% sequentially and 102% from a year ago. The AI Clouds, Industrial, and Enterprise group (ACIE) generated $40.3 billion, an increase of 25% from last quarter and 138% year over year.

While big tech is still the largest buyer of Nvidia's systems, it is no longer the fastest growing. That shift changes how the AI story should be read.

Image source: Nvidia.

The hyperscalers form Nvidia's foundation

The hyperscale bucket primarily revolves around the public cloud oligopoly: Amazon Web Services (AWS), Microsoft Azure, and Google Cloud Platform (GCP). These companies aren't buying Nvidia's products as a hobby. They buy them because GPU clusters have become a scarce resource supporting training runs, large-scale inference deployments, and the rented capacity that developers consume.

Despite designing its own Trainium, Inferentia, and Graviton silicon, AWS remains one of Nvidia's largest customers. The reason is simple: Nvidia's GPUs and CPUs play a critical role in how start-ups, labs, and enterprises rent Blackwell and Rubin chips from cloud providers without owning their own facilities packed with liquid-cooled racks.

Azure has bound itself to this same stack. Microsoft's Copilot suite, OpenAI-related training, and Azure-OpenAI services all rely on Nvidia systems. This structure is unique, as it makes Microsoft both a customer of and a distribution channel for Nvidia.

Google Cloud also designs its own custom silicon, called Tensor Processing Units (TPUs). Even so, Alphabet still buys enormous quantities of Nvidia hardware. One reason why is that many customers want access to CUDA, Nvidia's software ecosystem that runs on top of its GPUs. This makes the ability to move AI workloads across clouds much more efficient. For now, Google's TPUs cannot serve that type of demand on its own.

Taken together, these three cloud hyperscalers form the floor supporting Nvidia's data center operation. By signing multiyear capacity plans they essentially turn Nvidia's racks into an infrastructure-as-a-service empire. When revenue from the hyperscalers more than doubles in a year, it is evidence that the largest buyers see returns on procuring more accelerators. It is not evidence, however, that the GPU market only has three buyers.

How new adoption is becoming a big market for Nvidia

Analyzing the results from the ACIE customers is where Nvidia's story gets more interesting. Nvidia describes this category as AI natives, enterprises, sovereign customers, and the specialist clouds -- neoscalers like CoreWeave and Nebius Group -- that sit between a hyperscaler and a data center. During the second quarter, this mix of customers grew almost twice as fast sequentially as hyperscale. Over the last year, it grew even faster.

This is unique because it shows that AI budgets are not recycled through the same three offices. An AI cloud is an emerging kind of intermediary, one that buys Nvidia's chips, packs them into clusters, and leases capacity to companies that never need to negotiate with a chip designer directly. Industrial and enterprise buyers are even different. Manufacturers running digital twins on the factory floor, a bank assessing risk, or a government designing a sovereign cluster is far more complex than a corporation simply increasing its operating budget to rent incremental storage on AWS.

This distinction is important because skeptics seem to think that the AI revolution is confined to Amazon, Microsoft, and Alphabet. This gives bears an excuse to call AI a circular trade. Cloud giants spend on GPUs so they can sell AI services whose customers are none other than frontier model developers. But when ACIE customers outgrow the hyperscalers, Nvidia's roster of buyers becomes larger. This proves that demand is moving beyond a small cohort of platform businesses to a broader class that also needs generative models for production use cases, not just for platform differentiators.

Most importantly, rising sales from ACIE dampens the AI bubble argument. A bubble story has to explain why Nvidia's non-hyperscale book is accelerating. A new class of buyers means the capex supercycle is developing a second demand curve, and second demand curves are how infrastructure booms wind up creating new industries.

Where is the best place to invest to capitalize on AI infrastructure?

Don't get me wrong: this analysis is not to say that hyperscalers have become unimportant. AWS, Azure, and Google Cloud still remain the clearinghouses for much of the world's rented compute. Enterprise software names that are able to attach new AI-driven use cases to customer licenses will benefit if the ACIE demand is in fact real.

At the end of the day, hyperscalers still need Nvidia's systems to keep their AI products competitive. Neoscalers need GPUs and public clouds to exist to even have a viable business. Enterprises and sovereign governments rely on Nvidia's ecosystem spanning software, networking, and rack-scale design. This is all to say that accelerating ACIE growth shows how Nvidia's customer base is widening while the pick-and-shovel layer of capacity remains fairly concentrated.

Investors that fixate on the three cloud giants are stuck watching the first chapter of the AI narrative. Meanwhile, those watching enterprise software without asking who supplies the underlying compute are distracted by a third chapter that is still being written. The second chapter -- the one Nvidia's quarter actually highlights -- shines a light on the vendor that sells the AI factory to both. If AI adoption is truly broadening rather than looping, then Nvidia's scarce, high-margin systems will continue to be where evidence shows up first. In my eyes, that's what makes Nvidia such a no-brainer stock to buy and hold in the AI infrastructure era.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Amazon, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

Nvidia's (NASDAQ: NVDA) fiscal 2027 second-quarter earnings report was less of a quarterly update than it was a reminder that the company sits at the center of the artificial intelligence (AI) infrastructure build-out. Total revenue reached $96.2 billion, more than double the $46.7 billion posted a year ago and up 18% from the prior quarter. The more striking comparison, however, sits inside the underlying mix of Nvidia's sales.

The company's data center segment generated $89 billion alone. This single franchise now produces more sales than Nvidia's entire company did one year ago. These figures are proving that the hyperscaler capital expenditure (capex) boom is no longer an abstract backdrop. Cloud providers and AI infrastructure developers are adding capacity at full speed, and Nvidia is converting on that spend with unprecedented efficiency.

With that said, Nvidia's print did contain a quieter signal that may matter even more for the next name in the AI chip value chain: Micron Technology (NASDAQ: MU). Nvidia's management guided for a lower gross margin, and the explanation pointed directly at memory. This means that Nvidia is paying a premium for the stacks that sit beside its GPUs, making pricing power flow to memory suppliers like Micron.

Image source: The Motley Fool.

Nvidia's blowout quarter underscores its scale and dominance

Nvidia's numbers leave little room for understatement. Gross margin came in at 75%, essentially unchanged sequentially and up by 2.6 points from a year ago. Operating income rose 19% sequentially and 124% year over year to $63.7 billion. Meanwhile, net income was $59.7 billion, or $2.46 per share.

Within the data center book, hyperscale customers contributed $48.7 billion of revenue -- an increase of 102% year over year. It's clear that Nvidia is far from demand constrained, a distinction that showed up in both the beat against its own outlook and the size of the upcoming guide. Guidance is the other half of the print that I am more focused on. For the third quarter, Nvidia expects to generate revenue of $108 billion, plus or minus 2%. Of note, the company is assuming no compute sales to China.

Perhaps the most revealing line in the company's outlook is its gross margin. Generally accepted accounting principles (GAAP) and non-GAAP (adjusted) gross margin are both expected to be 74%, plus or minus 50 basis points. That is a full point below the margin Nvidia just delivered. Furthermore, management explained that gross margin is expected to bottom somewhere between 71% and 72% by the fourth quarter before recovering in fiscal 2028.

What is eating into Nvidia's margin?

Nvidia Chief Financial Officer Colette Kress was direct about the cause of the eroding margins. The step-down is clearly not a demand problem, nor is it a sudden collapse in Nvidia's pricing power with customers. Instead, it has to do with the rising cost of memory.

High bandwidth memory (HBM) is no longer a commodity add-on in chip clusters. Rather, it has quickly become a co-equal component of the accelerator package. Kress told investors that Nvidia's commitments for critical components increased from $119 billion last quarter to $279 billion, saying it was "primarily related to the procurement of memory."

I see Nvidia's margin cut as a positive signal for Micron. When a customer as large and as sophisticated as Nvidia accepts a lower gross margin outlook and cites memory prices as the culprit, it confirms the bottleneck is moving downstream into DRAM and HBM. Micron is one of a few producers that can supply the stacked memory AI platforms require.

Higher commitments from Nvidia are, by definition, higher realized average selling prices for memory vendors like Micron on the other side of the purchase order. Said differently, the same hyperscaler capex cycle fueling Nvidia's growth is now filling Micron's supply. The difference is that Micron is positioned to capture the inflation from component prices that Nvidia can no longer treat as a stable cost.

Nvidia's earnings performance is a preview for Micron

Investors no longer need to guess whether the AI infrastructure cycle is meaningful enough to reprice memory. Nvidia's lower margin guide just confirmed that it is. The dollars leaving Nvidia's gross margin are not disappearing. Instead, they accrue to companies like Micron, which are shipping HBM and adjacent DRAM into Nvidia's racks.

In my eyes, the prudent response is to watch Micron with the same intensity reserved for Nvidia. If the cost pressures Nvidia described are real, then Micron is positioned for the kind of print that should reset expectations and, potentially, the stock's momentum. The bull thesis around Micron stock is no longer a "hopium trade" attached to a historically cyclical name. Instead, it is quietly becoming an extension of Nvidia's own trajectory.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Micron Technology and Nvidia. The Motley Fool has a disclosure policy.

Nvidia (NASDAQ: NVDA) just told the market that it is willing to lock down more than a quarter-trillion dollars' worth of critical components to scale its data center empire. During second-quarter earnings, Nvidia CFO Colette Kress revealed that the company will be spending $279 billion on supply commitments over the next few years -- up from $119 billion only three months earlier. The reason for the jump is primarily about memory.

This figure is so large that it sounds less like a purchase order and more like an industrial policy. In an environment where everyone already knows high bandwidth memory (HBM) is scarce and expanding generative models need more of it, the size and speed of Nvidia's commitment are the real story.

Nvidia is not merely hedging by a quarter. It is pre-paying for the next several years of the artificial intelligence (AI) factory build-out so Blackwell systems and Vera central processing units (CPUs) can actually ship. Allow me to explain why.

Image source: Nvidia.

Nvidia's commitment redefines the AI supply chain

A budget of $279 billion is not a mere inventory buffer. It represents roughly three years of highly concentrated buying power pointed at the tightest part of the AI chip stack. Of this total, $92 billion is due in the remainder of fiscal 2027, followed by $87 billion and $88 billion across fiscal 2028 and 2029. Nvidia is effectively reserving the near-term memory market and not leaving it open until the next decade starts.

Image source: Nvidia Investor Relations.

The company's data center business is the key reason. Revenue from this segment reached $89 billion during second quarter fiscal year 2027, up 117% from a year earlier and on the way to a companywide guide of $108 billion for this quarter (fiscal third quarter).

Kress spoke about 70% total revenue growth in fiscal 2028, fueled by ongoing Blackwell shipments, scaling Vera Rubin as it reaches full production, and both pairing Vera CPUs with the company's existing graphics processing unit (GPU) architectures and selling them as a new stand-alone product. In essence, every extra rack, GPU, and CPU socket multiplies the HBM and server DRAM layered on top. Memory is no longer a line item seen as a commoditized, accessible solution. It has emerged as the bottleneck of AI infrastructure build-outs.

When a company with Nvidia's balance sheet increases supply commitments by more than double in a single quarter, it is signaling to producers that memory demand is real enough to underwrite new fabs. At the same time, it tells investors that the cost of this demand is going to show up in Nvidia's cost of goods before it fully materializes in higher selling prices down the road.

Understanding the AI memory tax

Scaling Blackwell and Vera is more than a silicon problem. More deeply, it is a packaging and memory problem. Each new generation of accelerators needs more HBM stacks, higher bandwidth, and tighter systems integration. Investors already know that memory prices are rising, though. What Nvidia's $279 billion supply and capacity budget does is prove that the company would rather absorb memory-driven inflation and guarantee future supply, instead of missing chip shipments.

The opportunity cost for Nvidia will be seen in gross margin. The company's gross margin was 75% last quarter and is guided to 74% this quarter. Management commented that a further dip into the low 70% range is realistic by the end of fiscal 2027. Higher average selling prices from DRAM and HBM are the most straightforward explanation for Nvidia's margin deterioration.

The bulk of Nvidia's memory spend will almost certainly land between SK Hynix (NASDAQ: SKHY) and Micron Technology (NASDAQ: MU). These two companies are at the center of HBM qualification for Nvidia's platforms. When Nvidia commits to a multi-year check this large, it is not spreading it evenly across a commoditized DRAM market.

Instead, Nvidia will concentrate its capital on two leading memory producers that it already knows can actually deliver the specialized components that make Blackwell and Vera perform as advertised. For Micron and SK Hynix, Nvidia's spending is not a cyclical restocking. It is a multi-year offtake that funds capacity additions these companies would have historically only hoped for.

Why AI memory stocks may be mispriced

Wall Street is already forecasting healthy revenue growth for SK Hynix and Micron over the next few years. However, the scale and duration of Nvidia's $279 billion supply order could easily drive higher-than-expected sales for both memory specialists. Taking this one step further, pricing power dynamics appear to be moving in favor of HBM and DRAM producers. This should fuel further earnings expansion for SK Hynix and Micron as Nvidia chooses to pay now rather than wait for supply to catch up with demand.

MU Revenue Estimates for Current Fiscal Year data by YCharts.

Despite this growth and compelling profitability dynamics, Micron and SK Hynix both trade at forward price-to-earnings (P/E) multiples around 6. I think this suggests the market is less comfortable owning the AI memory story as opposed to a known quantity such as Nvidia.

Two fears are keeping memory valuations in check: the notion that accelerating AI capex spend is a bubble that will burst, and the idea that memory is destined to follow its boom-bust cycle. In my eyes, Nvidia's data center results and the company's established $279 billion commitment mitigate both fears.

A bubble narrative has a hard time debunking $89 billion of data center revenue, a $108 billion next-quarter guide, and a 70% growth outlook for next year, all while the same company is simultaneously locking in three years' worth of memory solutions. Against this backdrop, cyclicality looks less realistic when the largest memory buyer in the AI landscape decides to pre-commit at this scale instead of risking spot orders that can vanish in a quarter.

This is why the better expression of an AI inflection may be investing in the memory suppliers rather than the platform owner. Micron and SK Hynix get more volume, pricing power, and visibility from Nvidia, all without having to defend margin erosion from the memory tax. All told, neither stock is priced as if a quarter-trillion-dollar customer order just reserved the next three years of inventory.

I think a prudent strategy is to buy Micron and SK Hynix together. They complement each other geographically and in product mix, all while sitting on the same Nvidia purchase order. While Nvidia remains the engine driving the AI revolution, these two memory producers are the fuel for the engine that the market still treats as a cyclical afterthought, for now.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Micron Technology and Nvidia. The Motley Fool has a disclosure policy.

Peter Thiel built his reputation by building and investing in companies that looked unfashionable until they became inevitable. Thiel co-founded PayPal in 1998 -- helping turn online payments from a futuristic curiosity into everyday e-commerce infrastructure. In 2003, he co-founded Palantir Technologies, the data analytics darling that now sits at the center of intelligence work with the U.S. government and its allies, as well as a growing number of Fortune 500 enterprises. Around the same time, Thiel wrote the first outside check into Facebook (now Meta Platforms). This investment ultimately became one of the defining venture bets of the internet era.

These chapters explain why smart investors monitor Thiel's moves closely. His investment philosophy is equally contrarian. In his book Zero to One, he argues that competition is for losers and that lasting value comes from building something no one else can truly copy -- essentially making your business a monopoly.

Perhaps unsurprisingly, Thiel has downplayed the loud narrative around artificial intelligence (AI). He sees the potential of the technology, but does not think it will lead to the absolute social transformation being promised by frontier labs.

The latest 13F filing from his hedge fund, Thiel Macro, puts this skepticism into quantifiable practice. After two consecutive quarters of holding nothing, the fund rebuilt a $419 million book during the second quarter. While Amazon is the largest position, roughly 33% of the portfolio now sits in three companies in the energy sector: Vista Energy (NYSE: VIST), Vistra (NYSE: VST), and X-Energy (NASDAQ: XE).

Image source: Getty Images.

Don't be fooled: Vista isn't an AI play

While energy consumption has become a popular discussion topic around AI infrastructure, I do not think data centers have anything to do with Thiel's position in Vista Energy. Instead, I see the position as more personal. Thiel Macro disclosed 1.2 million American depositary shares of Vista, worth $76 million at the end of the second quarter. This position makes up 18% of the hedge fund's disclosed book and roughly 1% of the company.

Vista is Argentina's leading independent energy producer in Vaca Muerta, a shale formation in Neuquén that holds the world's second-largest shale gas and fourth-largest shale oil resources. The company produces nearly 160,000 barrels of oil equivalent per day.

All told, Vista is really an oil and gas story rather than an AI play. In my eyes, Thiel's ties to the company are rooted more in geography and politics. While Thiel primarily resides in California, the billionaire purchased a mansion in Buenos Aires' Palermo Chico neighborhood earlier this year.

This isn't entirely surprising, as Thiel has met Argentina's President, Javier Milei, at the Casa Rosada and has previously spoken favorably of Milei's policies around spending, deregulation, and inflation. To me, Thiel's bet on Vista looks more like a wager that Milei's reforms will unlock the potential of a large, still-developing hydrocarbon basin.

Vistra: The AI link is more clear

During the second quarter, Thiel Macro built a $59 million position in Vistra -- representing about 14% of the fund. Of note, Thiel has owned Vistra stock in the past.

AI hyperscalers need carbon-free baseload power that does not wait for ideal weather conditions. Even a few hours of intermittent downtime wastes the capabilities of GPUs training AI models and running inference deployments. Nuclear brings this 24/7 optionality to the table. Vistra is a key enabler of AI workloads because the company operates nuclear plants in the Pennsylvania-New Jersey-Maryland (PJM) region.

It's important to highlight that Vistra has already locked in 20-year power purchase agreements with Amazon Web Services (AWS) for 1,200 megawatts from Comanche Peak and with Meta for more than 2,600 megawatts from its Ohio and Pennsylvania sites. Interestingly, Vistra also joined Nvidia, KKR, and the Kuwait Investment Authority in the Helix venture, a $10 billion partnership that aims to deliver nuclear power and data center infrastructure as a comprehensive package.

Given Thiel's prior history with Vistra and the company's relationships with a number of hyperscalers, I think this position represents his most obvious -- albeit still quite small -- positioning on the AI trade.

X-Energy: The newcomer at the intersection of nuclear and AI

X-Energy develops gas-cooled small modular reactors (SMRs). Prior to its IPO earlier this year, Amazon was a major backer and customer. While Thiel's post-IPO purchase is a discretionary expression of the same nuclear power-AI theme, smart investors must remember this is an extremely modest position for a billionaire.

In fact, X-Energy is the smallest line item in the entire Thiel Macro portfolio, comprising about 1% of the fund. Given its nominal scale relative to his other investments, I think this speaks volumes about Thiel's assessment of the AI narrative.

Taken together, these three energy stocks illustrate how Thiel is treating electricity -- not chips -- as the bottleneck around AI development. All told, Thiel's hedge fund is quite concentrated, and his positions and their relative sizes are consistent with a person who always prefers to own the enablers of mass adoption trends rather than the consensus trade.

Adam Spatacco has positions in Amazon, Nvidia, and Palantir Technologies. The Motley Fool has positions in and recommends Amazon, KKR, Meta Platforms, Nvidia, Palantir Technologies, PayPal, and Vistra. The Motley Fool recommends the following options: short September 2026 $47.50 calls on PayPal. The Motley Fool has a disclosure policy.

Space Exploration Technologies (NASDAQ: SPCX) began as a wager that rockets could be reused, launching into orbit could be cheap, and a private-sector business might be the vessel that carries civilization beyond a single planet.

While this origin story still sits at the center of the company's mission, SpaceX has changed some of its broader ambitions. Before its historic IPO, SpaceX absorbed xAI and subsequently folded rockets, Starlink, X (formerly Twitter), and a frontier generative AI model (Grok) into one vertically integrated moonshot. SpaceX CEO Elon Musk explained the rationale for this structure in a post on X: "You don't seem to understand that SpaceX will be worth more than the rest of Earth if we accomplish our goals."

This sentence is either a generational prophecy or hubristic salesmanship. Investors now have to decide which.

Image source: Getty Images.

A full analysis of the SpaceX IPO

When SpaceX opened on the Nasdaq in early June, shares came in at $150. By the end of its opening-day session, SpaceX boasted a market capitalization of $2.1 trillion -- immediately propelling it among the ranks of the world's most valuable companies.

The aftermath of SpaceX's debut has been less ceremonial. The stock peaked at $225.64 just four days following the IPO but has spent most of the summer learning how gravity works. By late July, shares dipped as low as $108 -- wiping out more than $1 trillion from SpaceX's market value. As of this writing (Aug. 26), SpaceX stock is changing hands around $138 -- essentially in line with the offering price of $135.

Measuring SpaceX's books against the worth of the planet

To be blunt, SpaceX's financial profile makes Musk's commentary look like a category error. In 2025, the company generated $18.7 billion in total revenue while incurring a $4.9 billion loss. While connectivity, led by Starlink, generated $11.4 billion in revenue and $4.4 billion in operating income, it was the only segment in the black.

The first half of 2026 has only accelerated the same pattern. Revenue in the first quarter was about $4.7 billion against a $4.3 billion net loss. The second quarter showed modest improvements across the board. Revenue surged 92% year over year to $7.8 billion, while net losses narrowed to $541 million.

The theme is that Starlink has become a real cash engine, but Starship and artificial intelligence (AI) infrastructure are enormous cash drains. Simply put, a company generating low-$20 billion of annual revenue, still burning cash, and spending like a sovereign enterprise is not, by any conventional methodology, a candidate to outvalue Earth. To close the gap, I think SpaceX would need to compound on three things simultaneously.

First, Starship would need to become so reliable and cheap that it makes space exploration a more recurring business. Second, Starlink has to continue scaling across consumer, enterprise, aviation, and maritime markets and would likely also need to enter telecommunications services -- all without compressing its average revenue per user (ARPU) into a commodity ditch. Lastly, SpaceX's AI capacity business needs to prove it can turn orbital compute from a slogan into legitimate contracted revenue that enterprise customers pay for at lucrative unit economics.

How should investors view an investment in SpaceX?

In my view, Starlink is worth owning in the ordinary sense: It has found product-market fit as evidenced by a growing subscriber base generating a healthy operating profit. The rest of SpaceX's equity story, however, is vulnerable to Musk's duration risk. Corporate governance is highly concentrated in Musk's voting power, while ballooning capital expenditures (capex) on emerging -- and somewhat still unproven -- businesses can overwhelm the cash war chest from the company's record listing.

In essence, a flailing Starship cadence or an uninspiring AI build-out that never recoups its cost of capital would leave SpaceX investors with nothing more than a premium-priced telecom-and-launch conglomerate, not a game-changing civilization utility.

Becoming the most valuable company in history could be imaginable if Orbital Compute actually works and reusable rockets collapse the cost of putting AI workloads into space. However, it is not the base case for a money-losing business generating just tens of billions of sales. At the end of the day, outvaluing Earth is not an investable forecast. Rather, it is a mission statement disguised as a price target.

The real takeaway is narrower and more harsh. SpaceX indeed offers a legitimate industrial complex with one profitable growth engine and two enormous call options attached.

With this in mind, I think SpaceX stock is only worth a position for investors who can stomach volatile drawdowns, tolerate a founder who has a history of answering questions with cosmology, and accept that shares could easily spend years looking expensive relative to its underlying books.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool recommends Nasdaq. The Motley Fool has a disclosure policy.

The S&P 500 (SNPINDEX: ^GSPC) has been on a multiyear rally for some time now. Annual returns reached 24% in 2023, 23% in 2024, and 16% in 2025. Meanwhile, the index has climbed another 12% so far this year.

These consecutive years of robust double-digit gains have produced a compound annual growth rate well above the market's long-term historical average. As the market flirts with record levels, it's clear investors are both confident and willing to bid stock prices higher on expectations of sustained growth.

While optimism fuels capital allocation, excessive exuberance can risk leaving valuations vulnerable if earnings or economic conditions disappoint. With that in mind, investors now face a classic tension: follow the momentum or employ some discipline to avoid overcommitment at elevated levels.

Understanding the market's current valuation profile

The cyclically adjusted price-to-earnings ratio, or CAPE ratio, is a valuation tool developed by economist Robert Shiller. The CAPE is calculated by dividing the price of the S&P 500 by the last 10 years of inflation-adjusted earnings. The idea is that by smoothing out cyclical fluctuations in corporate profits, CAPE readings provide a more accurate gauge of valuation that is less distorted by short-term swings than the more commonly used trailing price-to-earnings multiple. Elevated CAPE readings are historically associated with lower returns and a higher vulnerability to corrections, making the metric a useful cautionary warning sign.

S&P 500 Shiller CAPE Ratio data by YCharts

Since 2000, the CAPE ratio has had an annual average of roughly 28. This figure is well above the long-term historical average of about 18. The peak reading, however, occurred at the turn of the millennium. Back in 1999, the CAPE ratio reached 44, its highest reading on record.

This coincided with the height of the dot-com bubble. At the time, investors were recklessly pouring capital into internet companies, many of which generated little or no profit -- instead trading on visions of future network effects. The inevitable compression of earnings relative to soaring stock prices produced an extreme CAPE reading and set the stage for the subsequent multiyear bear market.

Why the current market may be different than the dot-com boom

On the surface, the current market environment shares some obvious similarities with the late 1990s. A transformative technology -- artificial intelligence (AI) -- has once again captured the imagination of investors. Just like 26 years ago, investors are driving unprecedented price appreciation in the technology sector. As a result, the CAPE ratio has risen into territory historically associated with elevated risk.

I think the underlying fundamentals of the current market dynamics diverge from the dot-com era in some important ways, though. Many of the companies at the center of today's AI boom generate substantial free cash flow, maintain durable competitive advantages, and are aggressively deploying capital into infrastructure such as data centers and advanced semiconductors. The earnings profiles of these companies are not prospective; rather, they are already materializing in current financials.

By contrast, the late 1990s were littered with numerous unprofitable internet businesses. Big tech leaders now operate across established industries and are embedding AI into products and services that drive measurable productivity gains across the broader economy. This makes the technology industry's tailwinds inherently more secular, resting on a firmer foundation rather than the speculative narratives that dominated the early days of the internet.

Image source: Getty Images.

How to prepare your portfolio when the CAPE ratio rises

While elevated CAPE readings frequently precede market corrections, they do not provide an exact calendar for when prices may reverse. Investors can still benefit from preparing without attempting to time an exact peak.

One approach is to trim a portion of gains accumulated in the strongest rallies while retaining core exposure to your highest-conviction holdings. These should be businesses that combine durable competitive moats, diversified revenue streams, and reliable cash generation. Rebalancing your portfolio toward these types of companies reduces concentration risk without abandoning the stock market entirely. Moreover, maintaining adequate cash and always employing a multiyear investment horizon cushions the impact of any interim sell-offs.

While history cautions that frothy valuations often lead to sharp drawdowns, the secular forces supporting today's market differ from those that accompanied the peaks of two and a half decades ago. Thoughtful portfolio construction helps acknowledge valuation risk while also allowing you to remain invested in quality stocks, ultimately offering a balanced response to the current market.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.

After Western Digital spun off Sandisk (NASDAQ: SNDK) last February, the company returned to being a stand-alone specialist selling NAND flash storage accompanying solid-state drives (SSDs) into three end markets: data centers, edge devices like PCs, phones, cars, and gaming consoles, and branded consumer storage solutions.

This all sounds quite boring... that is, until the workloads leveraging Sandisk's products changed. Training and running large language models does not stop at consuming GPUs. Generative AI also consumes high-speed storage for data lakes, model weights, and cache as inference deployments scale.

AI hyperscalers are increasingly expanding their capital expenditure (capex) budgets downstream (beyond chips), creating something of a supercycle in the memory market. Here's the thing: Most investors following the AI trade already know these industry dynamics. What they may not know, however, is how to price the scale of this move.

Over the past year, Sandisk stock has risen roughly thirtyfold. In 2026 alone, shares are up more than 500%, and that's after sliding about one-third from its peak back in June. The question is whether Sandisk's rally is finished. The case that it is not hinges on one number: $93.9 billion. Read on to learn why this figure is key to Sandisk's future.

Image source: The Motley Fool.

Sandisk's growth looks familiar

For the fiscal year ended July 3, Sandisk generated $20.2 billion in revenue, up 175% year over year. The company earned $73.76 GAAP earnings per share (EPS) after posting a loss in the year prior. On the surface, these figures don't reveal much other than that Sandisk is booming. But why? It's the company's revenue mix that tells the real story.

Edge remains Sandisk's largest business at $12.2 billion, up 195% year over year. This makes sense as AI-enabled PCs and phones are absorbing more flash. Sales from the company's Consumer division grew by a modest 29% to $2.9 billion. Meanwhile, Sandisk's data center business jumped 437% to $5.2 billion. During the fourth quarter alone, data center revenue nearly doubled sequentially to $2.9 billion and was up more than twelvefold from a year ago.

The rate of this growth is not merely a pedestrian imitation of Nvidia's (NASDAQ: NVDA) first surge during earlier phases of the AI revolution. During Nvidia's fiscal 2024, the company's data center revenue rose 217% to $47.5 billion. In the following year, Nvidia's data center business grew 142% to about $115 billion.

It makes sense that Sandisk's data center operation is smaller than Nvidia's in absolute dollars. A couple of years ago, the hyperscalers prioritized GPU procurement above anything else in the chip value chain. However, smart investors are starting to realize that initial waves of GPU demand fueled the current tailwinds supporting the AI memory landscape. Underneath the surface, investors can see that the percentage climb in Sandisk's data center business is already in the same neighborhood as Nvidia's early breakout.

Sandisk's new contracts are like Nvidia's chip architecture launches

The $93.9 billion figure referenced above is Sandisk's floor, not a lofty forecast. The company has signed New Business Model (NBM) agreements with eight data center and edge customers. These deals lock in committed bit volumes and mix fixed and variable pricing with floors and ceilings.

According to management, the NBMs run as long as five years and have a weighted average term of more than four years. At the end of the fourth quarter, Sandisk boasted $59.8 billion in remaining performance obligations (RPO), a figure that climbed to $91.1 billion after accounting for two post-quarter agreements.

The structure of these deals does for Sandisk what a new chip architecture used to do for Nvidia. When Nvidia first announced Hopper or Blackwell, the hyperscalers lined up almost immediately, providing the company with years of visible data center demand. Sandisk's new business contracts perform the same job without a product codename attached. In other words, these agreements convert Sandisk's historically cyclical price swings into a more defined backlog.

Sandisk stock is cheap relative to Nvidia's early breakout

Despite a stock price of nearly $1,500, the market is treating Sandisk stock like a cyclical memory name. Sandisk's price-to-earnings (P/E) ratio is around 20, while its forward earnings multiple hovers around 7.

SNDK PE Ratio data by YCharts.

In comparison, Nvidia was never this inexpensive on a forward basis in fiscal 2024 and 2025. Back then, Nvidia's forward P/E initially popped to around 30, but eventually sustained above 50 as the company captured the bulk of the initial AI infrastructure build-out.

NVDA PE Ratio (Forward) data by YCharts.

Sandisk is unquestionably the cheaper stock in this comparison, and by a wide margin. A company whose data center business is compounding at Nvidia-like rates, and that has nearly $94 billion of minimum contracted revenue, is being priced as if a downcycle is the base case. To me, this is the tell.

Sure, Sandisk's stock price has had a spectacular year so far, but its underlying valuation hasn't had its "Nvidia moment" yet. If the company's revenue mix continues shifting toward the data center segment, expansion in Sandisk's multiples should be inevitable.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Nvidia and Western Digital. The Motley Fool has a disclosure policy.

Peter Thiel gained early fame for being a co-founder of PayPal alongside Elon Musk. After selling that company to eBay, Thiel used his newfound fortune to become a venture capitalist. One of his earliest multibaggers came from investing in Facebook (now Meta Platforms) in 2004. Thiel later co-founded Palantir Technologies, the data analytics firm that has become a cornerstone of government and enterprise intelligence work.

Today, the serial entrepreneur manages capital through a hedge fund called Thiel Macro. According to the fund's latest 13F disclosure, in the second quarter, it opened a new stake in Amazon (NASDAQ: AMZN) -- acquiring 495,000 shares valued at roughly $118 million. This represents about 28% of the hedge fund's portfolio. This suggests that despite the stock already having a generational rise behind it -- with a 347,260% return since its IPO in 1997 -- Thiel still sees upside in Amazon.

That purchase raises an interesting question, though. Why would a contrarian thinker who is famous for seeking out monopolies suddenly invest in a company that faces intense competition on every front?

Image source: Amazon.

Thiel has a preference for monopolies, but Amazon faces stiff competition

Thiel has long argued that "competition is for losers." In his view, lasting value accrues to companies that can escape competition and establish durable monopolies through proprietary technology, network effects, economies of scale, or brand moats. Amazon fails this test across all of its major businesses.

In e-commerce, the company competes with Walmart's massive physical and expanding digital footprint in the United States, among other rivals. Meanwhile, Amazon remains virtually absent from the Chinese market, which is primarily dominated by local players. In cloud computing, Amazon Web Services (AWS) still leads in market share, but it is contending daily with Microsoft Azure and Google Cloud Platform, both of which are gaining ground.

The digital advertising space pits Amazon against the entrenched duopoly of Meta and Alphabet, while its Prime streaming service faces Netflix, Disney, and a crowded field of ancillary providers. Far from enjoying monopoly rents, Amazon operates in saturated markets where customers can switch providers, and rivals can undercut it on pricing at the flip of a switch.

What might Thiel see in Amazon?

Given that Amazon is not a monopoly and its aggressive capital spending on artificial intelligence (AI) has driven free cash flow into the negative, what might appeal to him about the tech giant as an investment is not immediately clear. What Thiel might be looking at is Amazon's infrastructure scale, which is giving it a more subtle competitive advantage.

AWS is the world's largest cloud platform and generates the bulk of Amazon's operating profits. After years of uninspiring growth, AWS' revenue gains are accelerating again as generative AI workloads surge. At the same time, AWS is designing a full-stack AI ecosystem featuring its custom Trainium, Inferentia, and Graviton chips, and expanding its suite of managed services. In addition, the company's large equity stake in Anthropic gives Amazon lucrative exposure to one of the world's frontier large language models (LLMs) in a way that doesn't require it to bear the full research risk alone.

It might be that Thiel is betting that Amazon's ability to supply the entire AI stack -- compute, storage, networking, specialized silicon, and model hosting -- will create compounding advantages that pure-play software or chip designers and manufacturers cannot match at scale. When viewed through this lens, the current pressures on its free cash flow can be seen as just the temporary price of locking in this strategic position while demand still exceeds supply.

Is Amazon stock a good buy?

Amazon is a tough stock to value. The lumpiness of its e-commerce business, combined with emerging services on so many other fronts, makes the company's net income quite unpredictable. For this reason, I personally do not love using the price-to-earnings (P/E) ratio to gauge the value of Amazon stock.

Instead, I prefer to look at the company on an enterprise-value-to-operating cash flow basis. Currently, Amazon's EV-to-OCF of 16.7 is near its lowest level since the start of the AI revolution. It's also notably under Amazon's P/E ratio of 21.

Fundamental Chart data by YCharts.

I see this disparity as quite meaningful for a company whose operating cash flow continues to expand even as its capex intensity accelerates. Against this backdrop, perhaps Thiel sees Amazon as a cheap way to own both the physical and software backbones of the AI boom.

While most investors focus on near-term free cash flow or competitive noise, Thiel appears focused on the multiyear optionality of an integrated ecosystem whose returns will continue materializing once ventures across new data centers and chips are fully utilized.

Amazon could be a reasonable buy for patient and disciplined investors who can tolerate ongoing elevated spending for a few more years. Ultimately, Amazon's combination of relative value and strategic positioning suggests Thiel could be early rather than reckless.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, and Palantir Technologies. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, Microsoft, Netflix, Palantir Technologies, PayPal, Walt Disney, and eBay. The Motley Fool recommends the following options: short September 2026 $47.50 calls on PayPal. The Motley Fool has a disclosure policy.

When it comes to the stock market, history may not repeat itself to a "T," but it often rhymes in ways that unsettle even the most confident investors. Markets tend to climb higher on waves of optimism, only for familiar pressures to reappear in new forms.

At recent prices, the S&P 500 (SNPINDEX: ^GSPC) sits around 7,670, a level that towers over the records of prior decades. After years of outsize advances, the index has delivered a roughly 12% gain this year. Yet the backdrop of this ascent includes lingering uncertainties about inflation, the Federal Reserve's next move on interest rates, and the ongoing Iran conflict.

Against this multi-year rally, the possibility of a correction is beginning to feel less like distant theory and more like a reality waiting to play out.

Image source: Getty Images.

How do we know the stock market is overvalued?

When it comes to valuation, investors can rely on two time-tested tools. The cyclically adjusted price-to-earnings (CAPE) ratio is a metric that was popularized by economist Robert Shiller. The CAPE ratio divides the current market price by the average of 10 years of inflation-adjusted earnings. By smoothing out isolated booms and recessions, CAPE readings more accurately determine whether investors are paying a premium for corporate profits.

The CAPE's long-run average hovers around 18, and readings above 30 have often preceded extended periods of single-digit or even negative returns. Currently, the CAPE ratio hovers around 41 -- among its highest levels ever recorded and within shouting distance of the all-time extremes seen during dot-com euphoria in the late 1990s.

S&P 500 Shiller CAPE Ratio data by YCharts.

Complementing the CAPE ratio is the Buffett indicator, which measures the total market capitalization of U.S. stocks relative to gross domestic product (GDP). As a rule of thumb, when the indicator climbs above 100%, stocks have historically been priced for perfection. Today, the Buffett indicator stands near 236%.

Both of these metrics are important because they drain out short-term noise and focus on the premium investors are currently assigning to future growth. While elevated readings in both ratios do not dictate any precise timing of a stock market correction, they reliably flag that the margin of safety is narrowing.

Will the stock market crash in 2026?

Elevated CAPE and Buffett readings have historically been followed by periods of digestion in which stock prices consolidate. With that said, there is no guarantee that a sharp correction must arrive within a specific time frame. Rather, these tools simply raise the probability that a meaningful pullback will occur at some point.

What makes the current environment unique is the unusual concentration of the market's gains in a handful of mega-cap tech companies: not just the "Magnificent Seven" -- Nvidia, Apple, Alphabet, Microsoft, Amazon, Tesla, and Meta Platforms -- but also Taiwan Semiconductor Manufacturing and Broadcom, which have also led the long-term rally and each sport valuations well over $1 trillion, like the others. These companies generate substantial free cash flow, dominate their respective industries, and continue to reinvest in transformative technologies -- namely, artificial intelligence (AI).

Unlike the speculative excess of the dot-com era, during which many unprofitable start-ups traded at fantasy valuations, today's market leaders are generally delivering tangible earnings and maintaining fortress balance sheets. This means the AI infrastructure wave rests on real productivity gains rather than pure narrative, for now.

While concentration risk is real, the underlying profitability of big tech provides a sturdier foundation than the froth between 1999 and 2000. A correction could -- and likely will -- still materialize, but the character of the current advances in the market differs enough that any eventual pullback won't resemble the exact cascading collapse of prior bubbles.

This investing strategy has never failed

Over every market cycle, the S&P 500 has always advanced -- recovering from recessions, wars, and valuation extremes. This long-run upward drift is perhaps the most reliable pattern in the stock market.

^SPX data by YCharts.

With this in mind, a practical response is pretty straightforward: Investors can buy a low-cost S&P 500 exchange-traded fund such as the Vanguard S&P 500 ETF. By committing to dollar-cost averaging -- investing a fixed sum at regular intervals -- regardless of the day's price, investors automatically buy exposure to the most resilient market vehicle of all time.

Over the course of a few decades, the compounding effect of reinvested dividends and gradual capital appreciation will turn modest contributions into substantial wealth. In time, investors will come to realize that drawdowns become opportunities to buy the dip. By remaining fully invested through the inevitable volatility, disciplined capital will capture the durability of the market's growth without the impossible task of timing every peak and valley.

Adam Spatacco has positions in Alphabet, Amazon, Microsoft, Nvidia, and Tesla. The Motley Fool has positions in and recommends Alphabet, Amazon, Apple, Broadcom, Meta Platforms, Microsoft, Nvidia, Taiwan Semiconductor Manufacturing, Tesla, and Vanguard S&P 500 ETF. The Motley Fool has a disclosure policy.

Recent 13F filings reveal that a flurry of top hedge fund managers bought the same "Magnificent Seven" stock during the second quarter: Amazon (NASDAQ: AMZN). These purchases by Stanley Druckenmiller, Peter Thiel, Seth Klarman, and David Tepper are more than coincidental portfolio adjustments.

I think this collective bet on Amazon underscores a recognition that the company is evolving far beyond its e-commerce roots into a platform business positioned for durable growth. What unites their thinking is a keen focus on emerging businesses that compound through technological leverage, cash generation, and competitive advantages that others in the big tech ecosystem struggle to replicate at scale.

Image source: The Motley Fool.

Converging on diversification

These hedge fund managers all share a preference for companies that blend defensive cash flow with offensive growth engines. Druckenmiller has long favored asymmetric upside in technology leaders while Thiel prioritizes network effects and monopoly-like businesses. Klarman tends to seek a margin of safety in undervalued assets while Tepper is always on the hunt for cyclical recoveries with secular tailwinds. Amazon satisfies each of these criteria.

Amazon's strength in artificial intelligence (AI) stems from its cloud computing division, Amazon Web Services (AWS). AWS provides a computational backbone for training and deploying AI models at a global scale. Beyond this, Amazon's logistics network, advertising platform, and subscription services (Prime) create a reinforcing ecosystem: Data from retail and enterprise customers improves ad targeting, fulfillment efficiency lowers delivery costs, and cloud infrastructure monetizes the resulting intelligence.

The company's diversification is the subtle strength here. Retail shoppers generate order volume, which helps Amazon build relationships with its customers. Advertising extracts high-margin revenue from these relationships while AWS delivers operating leverage. As a result, the company is able to reinvest in emerging efforts across robotics and custom silicon -- helping reduce dependency on external suppliers and manufacturers. Amazon's multi-legged stool reduces the risk of a single point of failure while amplifying returns.

The simultaneous buying by all these billionaire investors suggests they saw the last few months as a window where the market remained distracted by near-term retail pressures from inflation, ultimately obscuring the accelerating contribution from higher-margin, technology-enabled segments.

How should you value Amazon?

Amazon is a tricky stock to assess. While the company's price-to-earnings (P/E) ratio of 21 has compressed over the last year, this multiple could still appear elevated relative to more traditional industrial or retail peers. I think viewing Amazon through the lens of enterprise value (EV) relative to operating cash flow (OCF) paints a much different picture -- one in which the stock actually trades near historically low levels.

Operating cash flow better captures the real economic output of a business before the distorting effects of depreciation, stock-based compensation, and capital expenditures (capex). As the chart below illustrates, Amazon's ratio of enterprise value to operating cash flow currently hovers around 17, well below its 10-year average of 26.

Fundamental Chart data by YCharts

Since Amazon continually reinvests to expand capacity, its earnings can lag underlying cash generation. For this reason, the EV-to-OCF ratio better isolates the company's cash-generation potential once growth spending starts moderating or begins to yield higher incremental returns.

Against this backdrop, Amazon starts to resemble a hidden value stock: The market appears to be pricing it as a high-multiple hyperscaler while the cash metrics shown above reveal a business generating substantial liquidity at a discount relative to its long-term trajectory.

Savvy fund managers tend to prioritize cash flow and reinvestment runways over the optics of short-term earnings. I think the purchases discussed above imply that some money managers identified a temporary disconnect between Amazon's reported valuation multiples and the durability of the company's cash conversion.

What catalysts does Amazon have?

Amazon's horizon has several catalysts. Most obvious is AWS, which continues to expand its market share of enterprise AI workloads -- converting accelerating utilization into recurring high-margin revenue. Robotics initiatives should fuel productivity gains across Amazon's warehouses and last-mile delivery operations, helping lower unit costs as volume scales.

Moreover, the company's custom chips business aims to reduce reliance on external silicon designers and creates an additional product line beneath the AWS umbrella. Lastly, the anticipated IPO of Anthropic -- in which Amazon holds a significant equity stake -- should better crystallize the value of its generative AI ecosystem and provide a more visible validation of the company's strategic partnerships.

Taken together, these emerging opportunities support a compelling buy-and-hold thesis for Amazon stock. The company's unparalleled scale advantages, proven reinvestment discipline, and expanding technological moat suggest that disciplined capital will benefit from compounding effects over a long-term time horizon.

The recent moves by several fund managers highlight that the broader market may still be underappreciating the breadth and scale of Amazon's optionality. For investors willing to look past near-term volatility and focus on future cash generation, ongoing diversification, and AI-enabled operating leverage, the stock presents a unique intersection of growth and relative value.

Adam Spatacco has positions in Amazon. The Motley Fool has positions in and recommends Amazon. The Motley Fool has a disclosure policy.

The artificial intelligence (AI) semiconductor landscape is an interconnected web in which no single company operates in isolation. As the primary architect of the accelerators that power AI training and inference, Nvidia (NASDAQ: NVDA) sits at the center of this web. The company's earnings report this week will inevitably draw intense focus on Wall Street, yet the real narrative will likely extend beyond the company's own numbers.

Memory and data storage companies have become critical enablers across the broader digital ecosystem. During this earnings season, Apple CEO Tim Cook, Amazon CEO Andy Jassy, and Space Exploration Technologies CEO Elon Musk all highlighted the same pressure point: rising memory costs driven by soaring demand.

Investors monitoring Nvidia should keep a close eye on Micron Technology (NASDAQ: MU), Sandisk (NASDAQ: SNDK), and SK Hynix (NASDAQ: SKHY), as strong momentum in Nvidia's data center business will serve as an indicator of sustained demand for the specialized memory and storage components these companies produce.

Image source: Nvidia.

Nvidia's position in the AI chip value chain

Nvidia doesn't manufacture chips in the traditional sense. It designs graphics processing units (GPUs) and other processors, then relies on an ecosystem of partners to bring those designs to life. In the data center segment, Nvidia's powerful parallel processors form the computational backbone of hyperscale chip clusters.

However, each processor is only as effective as the high bandwidth memory (HBM) that feeds it data to process and the storage systems that manage these enormous data sets. This interdependence puts Nvidia in a coordinator role within the broader AI chip value chain. This means that the company's design decisions directly influence the technical requirements and volume forecasts for upstream suppliers.

Smart investors understand that when Nvidia reports robust growth in its data center segment, it is quietly confirming that big tech and enterprise customers are expanding their AI infrastructure at a rapid pace. In turn, these build-outs do not stop at GPUs -- they expand outward to the memory components that must keep pace with an accelerator's appetite for data.

The emerging theme across the AI semiconductor industry

Recent commentary from Cook, Jassy, and Musk has made the memory shortage impossible to ignore. The gap between the amount of DRAM and NAND flash memory that the memory manufacturers can supply and the volume that data center operators want is wide, and the rising prices for both underscore the mismatch between the growth in enterprise AI workloads and the capital-intensive, multiyear process of expanding fabrication capacity.

Nvidia's newer Blackwell and Vera Rubin chip architectures incorporate more HBM per unit than earlier designs. Meanwhile, hyperscalers are deploying these systems in ever-larger clusters. As a result, any discussion of data center momentum during Nvidia's earnings call will almost certainly highlight the limited availability and rising cost of the memory that stitches these systems together.

Nvidia CEO Jensen Huang has a reputation for providing candid color commentary about the state of the industry. The current memory environment offers him a natural opening to address ongoing supply dynamics. If he does, that would do more than restate what other executives have already said. Instead, it would better quantify the impact that the dynamics of the AI memory segment are having on the space from the perspective of the company whose products are driving the largest incremental demand.

Why Micron, Sandisk, and SK Hynix could respond to Nvidia's earnings

Micron, Sandisk, and SK Hynix occupy complementary positions in the memory and storage hierarchy that supports Nvidia's ecosystem. SK Hynix has established itself as a leading supplier of advanced HBM stacks that Nvidia integrates into its flagship accelerators.

Amid this boom in AI infrastructure investment, Micron has rapidly expanded its own high-bandwidth offerings while maintaining a broad portfolio of DRAM and NAND products across cloud, mobile, and automotive environments. Sandisk focuses more narrowly on NAND flash -- providing the high-capacity storage solutions and enterprise solid-state drives (SSDs) that hold data sets and model weights flowing through these systems.

The theme here is that when Nvidia's data center business is growing, the signal is subtle, but powerful: It indicates that customers are not only buying more GPUs, but also configuring these chips with a full complement of memory and storage required for production workloads. With that in mind, investors should be on the lookout for any qualitative remarks from Nvidia's management about customer deployment timelines and the mix of memory technologies being adopted.

Positive indications on this front should verify ongoing volume and pricing improvements for memory specialists. In addition, any acknowledgment of bottleneck constraints would further highlight the pricing power that memory suppliers currently enjoy.

The takeaway here is straightforward: Nvidia's upcoming earnings report and accompanying commentary should function as leading indicators for the memory supercycle. The three companies most closely aligned with AI-driven memory demand stand to reflect these signals in their own subsequent reports and stock performances as the AI infrastructure era matures.

Adam Spatacco has positions in Amazon and Nvidia. The Motley Fool has positions in and recommends Amazon, Apple, Micron Technology, and Nvidia. The Motley Fool has a disclosure policy.

In the earnings season that is now wrapping up, technology companies delivered a clear message regarding artificial intelligence (AI) spending. The five hyperscalers -- Microsoft, Amazon, Alphabet, Meta Platforms, and Oracle -- all either reaffirmed or boosted their capital expenditure plans for the year, pushing their combined planned outlay north of $700 billion.

The scope of that infrastructure spending is being driven by two interlocking forces: Insatiable demand for additional compute capacity as AI training and inference workloads explode, and a sharp rise in memory component prices that has made every new data center rack more expensive.

While most investors will be watching Nvidia's (NASDAQ: NVDA) fiscal 2027 second-quarter earnings report on Aug. 26, I'll have my eyes on Micron Technology (NASDAQ: MU).

Image source: The Motley Fool.

What is Wall Street expecting for Nvidia's Q2 earnings?

Wall Street analysts expect Nvidia to post roughly $92.1 billion in revenue and earnings per share (EPS) of $2.09 for the second quarter. That would amount to 97% revenue growth and nearly a doubling of profits year over year.

I see several factors that make a considerable earnings beat plausible. First, research analyst Vivek Arya from Bank of America notes that the four major cloud infrastructure platforms -- Azure, AWS, Google Cloud, and Oracle Cloud -- boast a combined backlog of $2.3 trillion. It's no secret that the hyperscalers remain capacity-constrained. Given this ongoing compute bottleneck, it's highly likely that demand for Nvidia's latest GPU and CPU platforms will remain robust. Moreover, Nvidia's management previously guided for gross margins to hold near 75%, underscoring the company's ability to maintain its pricing power.

How Nvidia's momentum flows directly to Micron

The impact of a strong performance from Nvidia shouldn't stop at that company's own stock price. Micron occupies a strategic position in the same AI chip supply chain. Nvidia's accelerators incorporate high bandwidth memory (HBM) stacks, and Micron is one of only three manufacturers qualified to supply Nvidia with these platforms. Micron's South Korean rivals, SK Hynix and Samsung, are the other two major high-end memory producers.

This means that every GPU that Nvidia ships requires a corresponding volume of advanced memory products. Because HBM production is increasingly being allocated to customers under multiyear agreements, any confirmation of accelerated GPU demand from Nvidia effectively tightens memory supply even further.

This, in turn, should fuel higher contract prices for both HBM and conventional DRAM -- translating into further revenue growth and margin expansion for memory suppliers like Micron. In essence, growth in Nvidia's data center chip shipments acts as a catalyst for Micron's pricing.

Why Micron stock could break out

After a parabolic rally earlier this year, Micron stock has spent the last month digesting its gains. This consolidation has left shares range-bound relative to pure-play GPU designers and other category-leading AI semiconductor stocks.

A decisive earnings beat from Nvidia would serve as an additional data point validating that the AI infrastructure supercycle remains intact, strengthening the value of memory as a key layer in the chip value chain. Considering that Micron's own HBM and DRAM output is already sold out and new capacity is still coming online, the narrative around the company should shift from "priced for perfection" to "still catching up with demand curves."

Investors waiting for a clearer catalyst to emerge may want to consider rotating some capital into memory names now, as I suspect a sharp rerating of these companies could be right around the corner. The combination of end-market strength from Nvidia and cloud hyperscalers, constrained supply across the industry, and a stock that has already absorbed considerable profit-taking creates a compelling setup for renewed upside once Nvidia's fiscal Q2 numbers hit the tape.

Bank of America is an advertising partner of Motley Fool Money. Adam Spatacco has positions in Alphabet, Microsoft, and Nvidia. The Motley Fool has positions in and recommends Alphabet, Micron Technology, Microsoft, Nvidia, and Oracle. The Motley Fool has a disclosure policy.

Marvell Technology (NASDAQ: MRVL) has delivered a standout performance among semiconductor stocks so far in 2026. The first catalyst arrived earlier this year in the form of a $2 billion investment from Nvidia -- bolstering a partnership that aims to deepen the technical collaboration between the two chip companies around interconnects and photonics.

This endorsement was amplified in June after Nvidia CEO Jensen Huang publicly asserted that Marvell could become the next trillion-dollar artificial intelligence (AI) chip company. With Marvell shares up more than 160% year to date and with its second-quarter earnings scheduled for Aug. 27, some investors may be wondering whether Marvell stock is still a buy.

Image source: The Motley Fool.

What is Wall Street expecting for Marvell's earnings?

Consensus estimates among analysts point to revenue of approximately $2.7 billion, which would amount to 35% growth year over year. Adjusted earnings per share (EPS) are expected to be $0.93, an increase of roughly 39%. Sustained growth at this scale implies Marvell has secured meaningful traction in custom application-specific integrated circuits (ASICs), high-speed networking, and optical connectivity from AI hyperscale operators.

Why timing your buys is a fool's errand

Attempting to time your buys right before or after an earnings release is not a sustainable investment strategy in the long run. The stock market prices a wide range of possible outcomes into shares ahead of such high-profile events, and the moves that follow an earnings report frequently are driven more by management's guidance and commentary than by the headline numbers themselves. Using a strategy of dollar-cost averaging avoids this noise because it spreads your purchases across multiple periods, reducing the impact of short-term volatility on your average purchase price.

How to build a position in Marvell stock

Given that Marvell's long-term value is tied to secular demand for AI infrastructure, periodic accumulation of its shares is the more rational approach. I think the prudent path in this case would be to wait until the market has fully digested Marvell's second-quarter results and management's outlook.

Moreover, smart investors may want to compare Marvell's growth rates to those of peer suppliers of network equipment and custom silicon, such as Broadcom. Examining the company in the context of its rivals should provide some insight into how Marvell's position in the AI chip value chain stacks up to those of its larger peers. Once the post-earnings dust settles, investors who remain optimistic about Marvell's role in custom silicon, optical networking, and data center connectivity can gradually build a position in the stock.

Spreading your purchases over the course of the multiyear AI infrastructure build-out allows you to avoid the binary risk of buying when shares may be overextended due to a single post-earnings reaction. In an environment defined by sustained capital spending on AI factories from big tech, steadily accumulating shares of a critical supplier is a more powerful and reliable tactic than attempting to catch the precise bottom or top around one quarterly report.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Broadcom, Marvell Technology, and Nvidia. The Motley Fool has a disclosure policy.

You probably know by now that a central question to the artificial intelligence (AI) narrative is not whether the technology will consume more memory but whether producers like Micron Technology (NASDAQ: MU) can convert ongoing shortages into a durable business model that compounds through the end of the decade.

For now, discussions primarily revolve around surging demand for high-bandwidth memory (HBM) and accelerating hyperscale capital expenditure budgets. While these talking points are necessary, they don't capture the full scale of the AI memory supercycle.

In my view, what matters more is a deeper shift that most investors are overlooking: Memory is ceasing to be a pure commodity and is quietly emerging as the binding constraint on the pace of further AI development.

Image source: Micron Technology.

Memory is a core feature of advanced computing

Each successive generation of AI models is hitting the same physical limit: Processors execute more parameters than the memory subsystem can actually feed. In this sense, Micron's nodes and advanced packaging are not simple incremental product upgrades -- they are the only practical way to keep extending AI's capabilities.

In the 2030s, the AI industry will almost certainly still be chasing denser memory stacks because demand from next-generation applications in robotics, autonomous vehicles, and agentic AI will be scaling alongside future breakthroughs in lithography.

Long-term supply contracts are helping Micron lock in volume and secure price floors for a meaningful fraction of its future output. I think these agreements are more likely to expand rather than expire as sovereign AI programs -- namely from SpaceX -- and edge deployments proliferate. As a result, Micron's earnings profile should become less volatile.

Micron has historically been (correctly) viewed and priced as a highly cyclical business, but against this evolving backdrop, investors may come to see it as an AI infrastructure compounder, and rerate the stock accordingly, granting it a higher earnings multiple.

Keep an eye on Micron's capital allocation

Another thing to monitor is the sheer volume of cash Micron generates once its production capacity expansions catch up to its contracted demand. Once this happens, Micron will be under less pressure to supplement its current $250 billion manufacturing build-out with outlays for additional fabrication capabilities. Instead, management will have a choice: Return capital aggressively through buybacks or dividends, acquire complementary intellectual property in packaging, or build internally and become even more vertically integrated.

Any of these paths should transform Micron's balance sheet from a cyclical buffer into a strategic fortress. Investors who only model peak-cycle earnings inherently miss the upside embedded in a growing cash position, as this capital can fund optionality in the product roadmap or simply reduce the share count -- both of which will amplify Micron's equity value even after the initial data center build-out.

What could an investment in Micron look like by 2030?

I think the best metric to use when analyzing Micron stock is the forward price-to-earnings (P/S) ratio because it requires two transparent assumptions: a level of normalized future earnings and a multiple the market will realistically assign once the memory cycle is better understood.

MU EPS Estimates for Current Fiscal Year data by YCharts.

In the chart above, investors can see that Wall Street analysts forecast Micron's earnings per share (EPS) will roughly double over the next year, but that growth will then begin to decelerate by calendar 2028. With these trends in mind, I've provided a scenario analysis below that shows a range of possible earnings figures and forward earnings ratios for 2030.

Given these assumptions, a $10,000 investment made in Micron today could be worth anywhere between roughly $8,400 and $36,600 by the end of 2030. I personally think the base case, which implies about 67% upside to current trading levels, is the most realistic. The bear case ultimately leads to capital erosion, a risk that could very well occur given the historically boom-and-bust nature of the memory industry. The big takeaway here is that an investment in Micron likely carries a spread that should not be overlooked.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Micron Technology. The Motley Fool has a disclosure policy.

New reports indicate that artificial intelligence (AI) lab Anthropic could file its S-1 by the end of the month. Below, I'll detail why Anthropic's public debut carries outsize implications for major backers like Amazon (NASDAQ: AMZN), whose growth is intertwined with the start-up's trajectory.

Image source: The Motley Fool.

Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »

Reviewing Anthropic's funding and IPO ambitions

Reports from Bloomberg suggest Anthropic is aiming for an initial public offering (IPO) that matches or exceeds Space Exploration Technologies' record raise from earlier this summer. Anthropic has already raised roughly $133 billion to date, most recently in a $65 billion Series H round that valued it at $965 billion. This near-trillion-dollar figure reflects Anthropic's explosive growth, with recent quarterly revenue surpassing $11.5 billion and an annualized run rate approaching $65 billion.

Why Amazon investors should pay attention to Anthropic's IPO

Amazon investors have good reason to monitor the Anthropic offering. Amazon has already invested $13 billion in Anthropic, with commitments for up to an additional $20 billion contingent on commercial milestones.

Beyond simple equity ownership, the partnership between Amazon and Anthropic runs deep through AWS. Anthropic uses Amazon's custom silicon for training and inference, including a massive deployment under Project Rainier that utilizes over 1 million Amazon Trainium chips, complemented by its Graviton processors.

Anthropic's Claude models power a number of features on Amazon Bedrock for more than 100,000 customers. Meanwhile, the company has pledged more than $100 billion in AWS spending over the next 10 years to secure up to 5 gigawatts of capacity.

Assessing Amazon's upside

Anthropic's S-1 will provide more precise disclosure around Amazon's ownership. With exact percentages known, investors can better model the position's value. Accurate knowledge of the value of Amazon's stake can help improve forecasts of AWS' revenue acceleration and operating margin expansion driven by Anthropic's compute demands. In turn, I think a fundamental rerating in Amazon stock could follow as investors gain a deeper understanding of Anthropic's influence on Amazon's position in the AI infrastructure supercycle.

Adam Spatacco has positions in Amazon. The Motley Fool has positions in and recommends Amazon. The Motley Fool has a disclosure policy.

Stanley Druckenmiller and Dan Loeb are two of Wall Street's most closely watched investors. Druckenmiller is a macro specialist who previously worked as a portfolio manager under George Soros. Today, he runs his own family office.

Loeb is the founder of the hedge fund Third Point. He's earned a reputation for being an activist investor and event-driven operator with a keen eye for identifying catalysts and pressing management for change.

Smart investors pay attention to Druckenmiller and Loeb because each has repeatedly demonstrated an ability to rotate capital into the next leg of growth before the consensus catches on. Perhaps unsurprisingly, their latest 13F filings show a shared pivot: Druckenmiller and Loeb both completely exited Broadcom (NASDAQ: AVGO) while building positions in Alphabet (NASDAQ: GOOG) (NASDAQ: GOOGL).

Below, I'll dig into why this rotation offers a valuable window into how two of the brightest minds on Wall Street are positioning for the next phase of the artificial intelligence (AI) revolution.

Image source: Getty Images.

Exiting Broadcom at a strategic juncture in the AI chip cycle

Broadcom is undoubtedly a central supplier to the AI infrastructure buildout. The company designs custom accelerators and also supplies high-performance networking silicon that links clusters of processors inside data center servers. Demand for these components has fueled Broadcom's revenue growth as AI hyperscalers race to expand capacity.

Yet the decision by both Druckenmiller and Loeb to fully exit their positions suggests they see better risk-reward elsewhere in the AI ecosystem. Despite some compression, Broadcom's valuation still prices in lofty expectations.

AVGO PE Ratio data by YCharts

Meanwhile, new GPU architectures from Nvidia and Advanced Micro Devices bring continued layers of competitive intensity to the equation. Lastly, capex cycles can shift quickly once capacity catches up with demand -- making Broadcom's long-term prospects questionable.

The thesis around Alphabet

Alphabet's appeal lies in its complete vertical integration. The company designs its own silicon, known as Tensor Processing Units (TPUs) and operates vast data center and fiber networks. Alphabet also develops AI models through its DeepMind lab under the Gemini family and distributes these capabilities at global scale across Google Search, YouTube, Android, and Workspace.

The theme here is that AI is not an add-on feature for Alphabet. Rather, the company's near-monopoly position in online search helps Alphabet weave AI features deeper into the fabric of every other major property the company owns. Search now surfaces AI Overviews to billions of users, while YouTube leverages generative tools for more personalized content recommendations and advertising.

Google Cloud is in a unique position because it sells both the infrastructure (TPUs) and enterprise services that feed usage directly data back into model improvement. This closed loop playbook creates compounding advantages that frontier model developers or merchant-chip suppliers struggle to replicate at scale.

Alphabet's AI strategy is producing impressive results. During the second quarter, the company reported total revenue of $119.8 billion -- an increase of 24% year over year. Google Cloud accelerated sharply, with revenue surging 82% to $24.8 billion. Meanwhile, backlog in the cloud segment reached $514 billion. Operating income for the company rose 30% to $40.8 billion, expanding Alphabet's operating margin to 34%.

Should you buy Alphabet stock right now?

Alphabet's vertical stack gives the company structural cost advantages, unparalleled distribution reach, and multiple monetization levers that competitors simply don't match. The company's recent earnings validate that the AI investments are generating both top-line acceleration and expanding margins -- translating into measurable traction rather than remaining an open-ended research experiment.

While blindly copying the trades of two billionaires is never wise, I think the logic behind this particular rotation is sound. Druckenmiller's and Loeb's simultaneous moves simply reinforce that two of the market's most successful capital allocators already identified the same opportunity.

Investors who believe the next wave of AI will reward platform businesses that own the full stack -- from model development, hardware, and software -- rather than component suppliers have a reasonable case for adding Alphabet to their portfolio.

Adam Spatacco has positions in Alphabet and Nvidia. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Broadcom, and Nvidia. The Motley Fool has a disclosure policy.

Right now, there are two companies dominating the artificial intelligence (AI) memory discussion: Micron Technology (NASDAQ: MU) and Sandisk (NASDAQ: SNDK). Analysis of which of these memory specialists deserves a place in your portfolio often centers on the same talking points: soaring demand for high-bandwidth memory (HBM) and flash storage.

What receives far less attention, however, is the architecture of each company's long-term positioning. Examining the lesser-discussed strengths and vulnerabilities of Micron and Sandisk suggests that the smarter choice is not to pick a single winner.

Image source: The Motley Fool.

Micron's edge in engineering flexibility

Micron's competitive advantage stems from the company's history of treating memory as a connected system rather than a commodity. While competitors continue chasing density milestones, Micron spent decades refining its DRAM and NAND portfolios so that the two can be co-optimized. This strategy is paying off as AI workloads increasingly demand both ultra-low latency and massive sequential throughput.

While this cross-pollination is not featured in headlines, smart investors understand that this approach allows Micron to prototype hybrid memory solutions faster than pure-play rivals such as Samsung or SK Hynix. As a result, Micron can respond swiftly to shifts in AI model architecture without waiting for its external partners to catch up.

Of course, investing in Micron does not come without risk. I'm not talking about the cyclical nature that has historically plagued the memory market, though. Micron's decision to diversify its memory portfolio across data centers, consumer electronics, automotives, and cloud computing can ultimately slow commercial rollouts when customers demand speed and access over perfection.

In an environment where AI labs and hyperscalers prioritize immediate availability over efficiency, Micron's wide reach can leave it temporarily behind more aggressive suppliers. Owning Micron stock requires patience for the company's measured scaling and conviction that this approach will lead to compounding effects or ultimately stagnate the business.

Sandisk's overlooked strength in firmware

Sandisk's narrative largely revolves around its focus on NAND flash. What receives almost no attention is the company's investments in firmware. This is important because these engineering innovations can extend the useful life of lower-cost flash. Given the pace AI infrastructure budgets are growing, Sandisk's ability to offer reliable performance from higher-end NAND offers a subtle cost advantage for hyperscalers.

Despite these advantages, Sandisk's primary liability is its concentration. Simply put, the company is tightly bound to NAND. This means investing in Sandisk requires an acceptance of more binary exposure to the trajectory of flash. Should there be an oversupply of these products, Sandisk's business becomes inherently devalued relative to a diversified peer.

Why pairing Micron with Sandisk makes sense

Another dimension to discuss is the way Micron and Sandisk interact with each other. Micron's broader footprint provides negotiating leverage with equipment suppliers and foundry partners. Meanwhile, Sandisk's narrow focus inherently pushes the performance envelope of next-generation NAND. In turn, this raises customer expectations for all suppliers across the memory value chain, including Micron.

This feedback loop is essentially invisible in price charts, yet it creates a strong mutual reinforcement in which investors should treat these two stocks as complements rather than substitutes. The obvious risk is that this interdependence goes both ways. If one company stumbles, the other's ability to bridge the gap will be constrained by its own specializations or scale limitations. This systemic fragility could amplify industrywide shortages in a more dramatic scenario.

Taken together, I think the lesser-spoken strengths and weaknesses around Micron and Sandisk ultimately point in the same direction. Micron offers architectural flexibility and wide distribution at the cost of more modest commercial scaling. On the other hand, Sandisk brings software-enabled longevity at the cost of product concentration.

Against this backdrop, I think owning both Micron and Sandisk does more than diversify cyclical exposure. Rather, a position in each stock brings more balance in a market that insists on crowning a singular AI champion. In my view, the more durable position is to hold the pair and let their complementary strengths compound throughout the AI infrastructure supercycle.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Micron Technology. The Motley Fool has a disclosure policy.

In February 2025, Western Digital spun off Sandisk (NASDAQ: SNDK) as an independent company. Shares began trading on the Nasdaq at around $35, and Sandisk stock has risen by a jaw-dropping 4,300% over the last 18 months. In 2026 alone, Sandisk has soared 568% -- making it the top-performing stock in the Nasdaq-100.

The company's rapid ascent has left some investors wondering whether the opportunity has already passed by. Let's explore the dynamics of the artificial intelligence (AI) memory market to better understand the drivers behind Sandisk's climb and assess if the rally can continue.

Image source: The Motley Fool.

Why is Sandisk stock up so much?

At the core of Sandisk's rise is a supply-demand imbalance amplified by AI infrastructure build-outs. Hyperscalers such as Microsoft, Amazon, Alphabet, Meta Platforms, and Oracle have collectively earmarked more than $700 billion in capital expenditures (capex) for 2026 alone. According to research from SemiAnalysis, big tech is expected to allocate 30% of its capex budget to memory solutions this year.

Sandisk is strategically placed in the NAND flash section of this AI spending. The company's enterprise solid-state drives (SSDs) are optimized for inference workloads and data lakes. This has fueled accelerated growth in Sandisk's data center segment, where trailing sales rose 437% over the last year to $5.2 billion.

What could propel Sandisk stock higher?

Several overlooked factors position Sandisk to sustain momentum. First, the company has locked in long-term "New Business Model" supply agreements with eight data center and edge customers.

These contracts carry a weighted-average duration of four years and guarantee a minimum of $93.9 billion in contracted revenue at floor pricing. This covers half of the company's fiscal 2027 bit supply and two-thirds of fiscal 2028. This level of visibility is unprecedented in a memory industry that long featured short-term contracts and boom-bust pricing.

Capital returns further reinforce the bull case. During the fourth quarter, Sandisk executed $4.5 billion in share repurchases and recently authorized an additional $15.5 billion buyback.

I think the most overlooked catalyst is Sandisk's joint venture with Japanese memory specialist Kioxia. This relationship helps Sandisk keep its capex at a minimal percentage of sales. In turn, the company can allocate more resources toward improving existing chip architectures rather than spending on additional manufacturing and factories.

Is it too late to buy Sandisk stock?

Wall Street's consensus estimates project Sandisk to more than double its earnings over the coming year. Sandisk trades at a forward price-to-earnings (P/E) ratio around 7. This is a steep discount compared to other leading chip stocks as well as the broader semiconductor industry average of 27.

SNDK EPS Diluted (TTM) data by YCharts

While memory stocks have historically traded at single-digit multiples due to cyclicality, Sandisk's multi-year backlog, growing free cash flow, and secular AI tailwinds justify a meaningful rerating. Against this backdrop, the current valuation profile still appears to embed lingering skepticism that the AI infrastructure cycle will revert, yet evidence of locked-in demand suggests otherwise.

In my eyes, Sandisk is far from priced to perfection. An investment in the company offers compelling upside thanks to the transformed economics of Sandisk's business.

Adam Spatacco has positions in Alphabet, Amazon, and Microsoft. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, Microsoft, Oracle, and Western Digital. The Motley Fool has a disclosure policy.

As founder of the hedge fund company Appaloosa Management, David Tepper has built a reputation for bold, often contrarian bets -- particularly in distressed assets. Investors monitor his moves closely because his track record includes legendary gains even during the toughest economic cycles.

New 13F filings reveal that during the second quarter, Appaloosa sold its entire position in Sandisk (NASDAQ: SNDK) while initiating a new stake in Broadcom (NASDAQ: AVGO). These moves offer a window into how Tepper is navigating the current chapter of the artificial intelligence (AI) revolution.

Image source: The Motley Fool.

Why selling Sandisk stock makes sense

Sandisk produces NAND flash memory, the high-speed storage technology that powers solid-state drives (SSDs). In the AI infrastructure era, these products have become indispensable because AI data centers require vast amounts of fast, dense storage to hold training data and inference workloads. AI-driven demand has fueled NAND prices sharply in recent quarters, shifting Sandisk's revenue mix toward enterprise data center customers. Revenue and margins are expanding dramatically as sales surge and pricing power returns to the storage specialist.

Filings show that Tepper initiated the Sandisk position during the first quarter of 2026 and subsequently exited completely sometime in the second quarter. While exact entry and exit prices are private, the timing alone implies substantial gains.

SNDK data by YCharts.

Taking profits after such a parabolic move signals discipline. Smart investors understand that memory and storage markets are cyclical, meaning Sandisk's valuation may already price in much of the near-term AI surge. By securing gains after a single quarter, Tepper reduces exposure to unwanted volatility while freeing capital for other opportunities.

David Tepper. Image source: Getty Images.

Why Broadcom might appeal to Tepper

Broadcom designs and supplies semiconductors and infrastructure software. Over the last few years, the company has focused on custom AI accelerators known as XPUs. These chips are designed alongside hyperscale customers, rather than being sold as general-purpose products.

Broadcom makes Google's Tensor Processing Units (TPUs) as well as Meta Platforms' MTIA chips. The company also expanded relationships with OpenAI, Anthropic, and Apple, further diversifying the customer base. Broadcom also supplies high-speed networking silicon that connects large clusters of these accelerators, giving the company exposure to both the compute and interconnect layers of the AI chip stack.

Despite trading at a lofty valuation based on price-to-earnings (P/E) and forward earnings multiples, Broadcom remains an attractive opportunity given its growth potential. AI semiconductor revenue is scaling rapidly, with management guiding for more than $100 billion in this segment alone by fiscal 2027.

AVGO PE Ratio data by YCharts.

Robust operating margins, a wide competitive moat in custom silicon, and a backlog of multi-year orders support Broadcom's premium valuation profile. For an investor like Tepper, the combination of visible multi-year growth, sticky customer relationships, and market leadership in a shift toward specialized AI hardware may outweigh rich multiples.

Should investors follow Tepper's lead and buy Broadcom stock?

Tepper's rotation away from a high-flying memory name into a broader AI infrastructure stock reflects disciplined capital allocation rather than an endorsement of any specific stock. Smart investors can reasonably take note of his logic: Book gains after an abnormal rally and redeploy capital into a company with deeper competitive advantages and potentially longer runway.

While Sandisk's recent pullback from its June peak has so far validated Tepper's exit, Broadcom's elevated valuation leaves little room for error. Against this backdrop, simply following Tepper's trade carries risks.

In my eyes, the more prudent approach is to study his underlying thesis rather than copying the ticker symbols. Investors who share Tepper's conviction in sustained AI capex spending from big tech may find Broadcom a compelling long-term holding. With that said, those who prefer pure play memory exposure could still find opportunities in the sector even after volatility subsides.

Regardless, Tepper's latest moves underscore a timeless lesson: Even the best investment ideas eventually require some harvesting, and the strongest portfolios always adjust to the market rather than remaining static.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Apple, Broadcom, and Meta Platforms. The Motley Fool has a disclosure policy.

After a stint on the high-yield desk at Goldman Sachs, David Tepper launched the hedge fund Appaloosa Management in the early 1990s. Over the last couple of decades, Tepper has generated an average annual return in the mid to high 20% range -- highlighted by an outsize performance in 2009 after he bought distressed bank securities near their lows during the financial crisis.

Combined with his ownership of the Carolina Panthers football team, Tepper's fortune has made him an investment personality whose moves are dissected for clues about the market's direction. During the second quarter, Appaloosa's 13F filing with the Securities and Exchange Commission showed that the firm fully exited its position in Lyft (NASDAQ: LYFT) while simultaneously adding more than 1.3 million shares of its ride-hailing rival, Uber Technologies (NYSE: UBER). Uber is now one of Appaloosa's five largest positions, representing about 7% of the portfolio.

Investors watching Tepper closely see this transaction as more than a simple rotation. Rather, it reflects a calculated judgment about relative competitive strength and long-term value creation in an intense ridesharing and delivery landscape.

David Tepper. Image source: Getty Images.

Breaking down Tepper's Lyft trade

According to filings, Appaloosa initiated its stake in Lyft during the first quarter of 2024, buying 467,618 shares. Throughout the rest of the year, its position grew to 13.5 million shares. While Tepper held the stock for roughly two years, his fund steadily pruned the position throughout 2025 and fully exited during the second quarter of this year.

I think the decision to exit was influenced less by any problems at the company and more by a broader desire for sharper focus in the industries in which Lyft operates. The company continues to post respectable growth in rides and gross bookings, but it remains a much narrower service provider whose scale lags that of Uber.

Analyzing Uber's business results

Uber and Lyft compete in overlapping markets, yet Uber's more-diversified platform and stronger financial momentum make it a more compelling long-term holding. During the second quarter, it reported gross bookings of $58 billion, up 24% year over year. The number of trips grew 18% to 3.9 billion, driven by robust growth in monthly active platform consumers (MAPCs).

These performance metrics translated to 33% growth in earnings before interest, taxes, depreciation, and amortization. Free cash flow for the quarter totaled $2.8 billion, lifting Uber's trailing-12-month free cash flow above $10 billion for the first time. This performance proves Uber commands impressive operating leverage across its mobility and delivery segments, both of which are supported by the company's expanding higher-margin advertising services.

Should you buy Uber stock right now?

The consensus price target for Uber among Wall Street analysts is $101, implying roughly 30% upside to current trading levels. This disconnect between the share price and Wall Street's forecast can largely be explained by persistent anxiety over the disruption promised by autonomous vehicle (AV) fleets.

Expanding services from Alphabet's Waymo and Tesla's Robotaxi have come with a perception of increased competitive pressures. This has resulted in significant multiple compression relative to Uber's historical valuation profile. Nevertheless, management is quietly scaling up its own AV partnerships and targeting several cities for launches over the coming quarters.

UBER PE Ratio data by YCharts; PE = price to earnings.

Uber's network effects, global footprint, and proven ability to convert rider and order volumes into expanding margins provide a durable foundation that robotaxi fears shouldn't erode overnight (if at all). The combination of accelerating free cash flow, an attractive valuation, and its model for adapting to embrace autonomous vehicles creates an asymmetric opportunity most investors appear to be overlooking.

Adam Spatacco has positions in Alphabet and Tesla. The Motley Fool has positions in and recommends Alphabet, Goldman Sachs Group, Lyft, and Tesla. The Motley Fool recommends Uber Technologies. The Motley Fool has a disclosure policy.

As the founder of Appaloosa Management, David Tepper built his reputation through a keen eye for distressed assets, including debt instruments trading at steep discounts, and deep-value stocks that most other investors overlooked amid temporary turmoil.

Tepper's willingness to embrace situations that others flee has helped him produce outsized returns over the long run. This makes his latest combination of bets between CoreWeave (NASDAQ: CRWV) and Meta Platforms (NASDAQ: META) particularly noteworthy for investors tracking the artificial intelligence (AI) infrastructure supercycle.

Image source: The Motley Fool.

The rise of neoclouds

Recent 13F filings show that Appaloosa initiated a position in CoreWeave stock during the second quarter, acquiring 1,078,248 shares. CoreWeave operates as a neocloud provider, delivering high-performance GPU capacity tailored to AI workloads. In an environment where hyperscalers are struggling to meet explosive AI-driven demand for training and inference deployments, neoclouds like CoreWeave bridge the gap by offering dedicated, scalable compute infrastructure optimized for Nvidia hardware and adjacent systems.

During the second quarter, CoreWeave's backlog grew by more than 246% year over year to $104 billion. Of note, this figure does not include new commitments of $25 billion signed shortly after the quarter ended. CoreWeave's customers include frontier AI labs, hyperscalers, and an expanding roster of enterprises. Revenue more than doubled year over year, reflecting the pace of rapid capacity deployment.

With that said, the company's balance sheet reveals a potential blemish that has left some on Wall Street uneasy. CoreWeave's total debt -- including operating leases -- stands north of $50 billion.

CRWV Total Long Term Debt (Quarterly) data by YCharts.

This capital from its debt sales is being used to finance the capital expenditures of its aggressive data center build-out. However, the company's leverage profile raises questions about possible refinancing needs in a higher-interest-rate environment. The combination of heavy borrowing and pressures to free cash flow has some investors concerned about the company's execution risks.

Why Tepper's Meta position looks interesting

During the second quarter, Appaloosa also increased its stake in Meta Platforms by 55%, adding 238,500 shares. This move looks particularly interesting because Meta is reportedly exploring the idea of leasing some of its excess AI compute capacity to third parties. This strategy could place the social media giant and hyperscaler in direct competition with neocloud operators such as CoreWeave.

While details remain sparse, that new revenue stream could help offset the costs that Meta is incurring to expand its data center footprint. What's even more interesting is contemplating the prospect of a well-capitalized CoreWeave customer -- Meta -- turning into one of CoreWeave's rivals.

The natural thought is that if hyperscalers start to fill more of the cloud capacity demand that they have previously outsourced to neoclouds, it could shrink the addressable market for those pure-play neocloud providers. Against this backdrop, Tepper's simultaneous commitments to both Meta and CoreWeave stand out against the potential friction.

How should investors view Tepper's positioning?

In my view, Tepper has built a unique hedge within the neocloud space by pairing his Meta and CoreWeave positions. Valuation analysis suggests that both companies can be viewed through a similar lens. Over the last several months, the forward price-to-sales (P/S) multiples of Meta and CoreWeave both compressed. Of note, Meta is also trading at a discount to its recent levels on a forward earnings basis.

CRWV PE Ratio (Forward) data by YCharts.

These trends could suggest that the market is already factoring much of CoreWeave's leverage risk into the share price while remaining skeptical over Meta's elevated capital outlays. I find downward pressure on both stocks interesting, as each company has consistently demonstrated robust growth throughout the AI boom.

Through his positions in CoreWeave and Meta, Tepper gains exposure to a market leader in the specialized neocloud landscape alongside a technology blue chip that is building a more three-dimensional AI ecosystem.

Whether this structure proves profitable will hinge on the companies' execution and the persistence of compute scarcity. If CoreWeave's remaining performance obligations (RPO) convert at a pace that allows it to retire debt at a meaningful pace, this position could bring asymmetric gains. At the same time, Meta's scale and diversification beyond advertising offer some balance should the valuations of pure-play neoclouds experience more pressure.

Tepper's track record of taking advantage of undervalued opportunities supports his strategy with this pairing, reflecting calculated opportunism rather than simple momentum chasing. In the long run, Appaloosa should be positioned to benefit regardless of how the competitive dynamics in the AI infrastructure space ultimately play out.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Meta Platforms and Nvidia. The Motley Fool has a disclosure policy.

When it comes to social media, most people probably think first of Facebook or Instagram -- two flagship properties owned by Meta Platforms. Not too far in the distance behind them, however, is rival Reddit (NYSE: RDDT). Reddit operates a vast network of chat forums where people discuss everything from their hobbies to global news events.

The company's core business model relies on advertising revenue, selling targeted placements to brands that value the platform's engagement. In recent years, Reddit has also started to monetize its enormous library of human conversations through data licensing. This has created a secondary revenue stream that aligns with the artificial intelligence (AI) boom.

The latest quarter's batch of 13F filings shows that Duquesne Family Office, the private firm of legendary investor Stanley Druckenmiller, initiated a position in Reddit during Q2 -- buying 56,550 shares. Given Druckenmiller's ability to spot asymmetric growth opportunities, I was curious about what attracted him to Reddit, and whether the internet darling is a potentially sound investment opportunity right now.

Image source: The Motley Fool.

Analyzing Reddit's performance as a public company

Reddit debuted on the New York Stock Exchange in March 2024. While the company priced its shares at $34, they actually opened closer to $47. In the months that followed, Reddit stock climbed steadily as the company demonstrated accelerating revenue growth and an improved profitability profile.

By September 2025, Reddit reached an all-time high price of roughly $283. Over the last year, however, shares have sold off considerably. As of this writing (Aug. 19), Reddit trades at around $155 per share -- a decline of more than 40% from its high.

So far this year, the stock is down about 32% -- lagging the S&P 500 (SNPINDEX: ^GSPC) by a wide margin. Reddit's underperformance is a reflection of investor concerns about the pace of the company's data-licensing deals and questions around the sustainability of its advertising sales growth given the competitive landscape.

Reddit's AI ambitions are interesting

Reddit possesses one of the richest archives of real human conversation online. This data is highly valuable to AI developers seeking high-quality training data that captures natural language, cultural nuance, and specialized knowledge. Back in 2024, Reddit struck multiyear licensing agreements with OpenAI and Google, granting each company access to the platform's content for model training. It's important to note that recent reports suggest there are new renegotiations ongoing between Reddit and Google, as Reddit has found this relationship cannibalizing some of its organic traffic.

While AI-driven revenue only represents about 5% of Reddit's total quarterly sales, that's a high-margin business line with meaningful room to expand as more AI labs seek licensed rather than scraped data. For an investor like Druckenmiller, Reddit's dual identity as both an advertising platform and an emerging AI data supplier offers a compelling combination of revenue and cash generation supplemented by future optionality.

The investment case for owning Reddit stock

Perhaps the most immediate catalyst for Reddit arrived last week when the company announced it was being added to the S&P 500. Inclusion in the index signals that the company has achieved sufficient scale, liquidity, and financial maturity to place it among the ranks of America's largest public companies.

Exchange-traded funds and mutual funds that track the S&P 500 are now required to hold Reddit stock, inherently creating structural demand that can support the stock price over the long run. More importantly, membership in the large-cap index underscores Reddit's evolution from an online forum into a more sophisticated, durable business capable of sustained growth and profitability.

Druckenmiller's decision to build a position in Reddit appears to have been well timed, capturing both the index-driven inflows and the longer-term potential of a platform deepening its AI relevance while expanding its legacy advertising footprint. Investors who share Druckenmiller's conviction in secular technology trends may consider following him into the stock.

RDDT PE Ratio data by YCharts.

Reddit's pullback throughout this year has left it trading at a more attractive price-to-earnings ratio relative to prior periods. The combination of proven revenue momentum, S&P 500 status, and expanding AI licensing creates an attractive risk-reward profile for those willing to look beyond the stock's near-term volatility.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Meta Platforms and Reddit. The Motley Fool has a disclosure policy.

As a macro trader whose career has featured numerous bold moves, Stanley Druckenmiller is one of the most respected figures in modern investing. As the former portfolio manager of George Soros's Quantum Fund, he helped orchestrate the 1992 short of the British pound -- a trade that generated a cool billion-dollar profit.

He later turned Duquesne Capital into an investing machine that delivered average annual returns near 30% with no down years before closing the fund in 2010 and shifting his priorities to his own family office. Investors pay close attention to Druckenmiller because of his uncanny pattern-recognition skills that have repeatedly identified inflection points ahead of the crowd.

Stanley Druckenmiller. Image source: Getty Images.

This makes his recent purchase of Tesla (NASDAQ: TSLA) call options -- a bet on a rising share price -- particularly interesting. Even as the stock plummets more than 20% so far this year and continues to trade at valuations that would make most traditional investors run for the hills, I can't help but think Druckenmiller sees something in Tesla that others are missing at the moment.

Image source: The Motley Fool.

Analyzing Tesla's valuation and the narrative behind it

Tesla commands a market capitalization of almost $1.4 trillion. This translates into a trailing price-to-earnings (P/E) ratio of more than 300, while its forward-earnings multiple remains elevated in the 190 range. Tesla's price-to-sales (P/S) sits near 11.

TSLA PE Ratio data by YCharts.

By comparison, established auto manufacturers like General Motors and Ford trade at single-digit forward-earnings multiples and P/S ratios of less than 1. Even pure play electric-vehicle (EV) peers like BYD sport more modest valuation multiples, typically less than 20 times forward earnings and about one times sales.

Tesla's valuation premium exists entirely because investors are assigning substantial value to two nascent artificial intelligence (AI) ambitions: the Optimus humanoid robot and the Robotaxi driving network. In effect, investors are currently paying for the possibility that these efforts will generate substantial profits and network effects far beyond the economics of selling cars.

Where do Optimus and Robotaxi currently stand?

Optimus remains firmly in the prototype phase. Although factory lines at Tesla's Fremont, California, facility are being installed after the decommissioning of older EV model production, build-outs for training data collection are still limited. For now, the robot is not a commercial product generating revenue.

Meanwhile, Robotaxi has finally moved beyond demonstration but has yet to deliver the disruptive scale Elon Musk long promised. According to Tesla's second-quarter earnings report, cumulative paid miles for Robotaxi climbed past 2.4 million. However, smart investors discovered that the pace of Robotaxi's growth actually flattened in recent months.

Image source: Tesla Investor Relations.

Quarterly additions stalled near 900,000 miles -- slipping even as the Robotaxi service expanded into additional metropolitan areas. Moreover, much of the active fleet still operates with safety monitors in the cars. What Musk once characterized as an imminent fleet of a million autonomous vehicles looks more like a carefully controlled pilot whose adoption has not accelerated with the optimistic rhetoric.

Why Druckenmiller's Tesla position could still pay off

Despite Tesla's frothy valuation and the shortfalls of Robotaxi so far, Druckenmiller's decision to buy call options might still be defensible. I should note that his 13F filings do not specify which option chain Druckenmiller specifically bought. This means that his calls could be short-dated or stretch well into the future. Either way, I think the position will be profitable.

My reasoning revolves around growth from Tesla's autonomous driving software. My suspicion is that Druckenmiller is ignoring any noise around Robotaxi and is counting more heavily on the scale and margin potential of Tesla's full self-driving (FSD) platform.

During Q2, active FSD users reached 1.48 million -- a 56% increase year over year. More than 55% of new deliveries in North America now include the FSD feature, which requires a subscription. This is important because the recurring revenue nature of FSD translates into software-like margins for Tesla, helping offset capital-intensive initiatives like Optimus and Robotaxi.

More importantly, the expanding base of FSD users feeds Tesla's proprietary data library. This helps the company accumulate real-world driving miles at a scale no competitor has yet to match, positioning Tesla as a potential first-mover with a durable competitive advantage in autonomous driving.

While I suspect Tesla stock will remain volatile for the time being, smart investors are looking beyond the company's AI vision and focusing more clearly on its established product lines. With FSD subscriptions fueling Services revenue, I think more investors will come to realize Tesla is quietly evolving beyond an EV manufacturer and finally -- albeit slowly -- becoming the tech-enabled platform long promised by Musk.

Adam Spatacco has positions in Tesla. The Motley Fool has positions in and recommends Tesla. The Motley Fool recommends BYD Company and General Motors. The Motley Fool has a disclosure policy.

Over the past week, a flurry of 13F filings have revealed the portfolio moves institutional investors made during the second quarter. The shifts from one such billionaire investor, Stanley Druckenmiller, captured my attention.

Druckenmiller's Duquesne Family Office fully exited Broadcom (NASDAQ: AVGO) while initiating a new position in Alphabet (NASDAQ: GOOGL). I find this move particularly interesting because Alphabet relies on Broadcom to help design its custom Tensor Processing Units (TPUs), the specialized chips that power a portion of Google's artificial intelligence (AI) infrastructure.

By favoring the company that is aggressively spending rather than the supplier receiving those dollars, Druckenmiller is signaling a nuanced view of the AI infrastructure trade. Rather than quietly riding the wave of ongoing chip demand, he may be looking further downstream toward the platforms tightly integrating these accelerators with the end-user experience.

Image source: The Motley Fool.

The case for selling Broadcom stock

One reason for selling Broadcom stock hinges on valuation and the evolving nature of the AI semiconductor cycle. Despite some valuation compression in recent months, Broadcom's price-to-earnings (P/E) and forward P/E are still frothy. The broader semiconductor industry trades at a 43 P/E and about 26 times forward earnings. Against this backdrop, it could be argued that Broadcom is still pricing in much of the near-term upside from hyperscaler contracts.

AVGO PE Ratio data by YCharts

While Broadcom remains a critical partner for large buyers, its AI segment leaves the company exposed to external spending decisions from big tech. Druckenmiller has a history of moving early when the risk-reward equation changes. By taking profits in Broadcom now, he may have concluded that easy gains from the pure AI infrastructure layer have been captured and that incremental returns will be harder to achieve.

There is also the simple discipline of portfolio concentration. What I mean by that is holding both the buyer and the seller of the same chips creates unnecessary correlation risk if the broader AI spending narrative decelerates.

What makes Alphabet an attractive investment?

Alphabet possesses massive data assets through Google search and YouTube, world-class research talent (DeepMind), and a proprietary hardware stack. Its TPUs are not merely cost-saving tools; rather, they are tightly integrated with the software and services that generate a portion of Alphabet's AI-driven revenue.

Owning the company that both designs these chips and deploys them across an AI ecosystem spanning search, cloud computing, and consumer products gives investors deep exposure to Alphabet's entire value chain rather than a single link.

Recent acceleration in Google Cloud growth has proven that the company's capital expenditures are translating into higher-margin recurring revenue and widening operating profits. All told, Alphabet quietly offers a leveraged play on AI without the volatility of pure play chipmakers themselves.

Druckenmiller isn't the only billionaire who likes Alphabet

The most compelling aspect of investing in Alphabet is the degree of the company's vertical integration. By controlling the design of its custom accelerators, the software that runs on them, and the global distribution platforms that monetize the resulting intelligence, Google has built an extremely profitable self-reinforcing system that is difficult for competitors to replicate at scale.

GOOGL Net Income (TTM) data by YCharts

This advantage helps explain why Alphabet has swiftly emerged as a core holding at Berkshire Hathaway. Berkshire appears to be treating Alphabet less like a speculative technology name and more like a durable franchise. During the second quarter, the investment conglomerate plowed $17 billion into Alphabet alone.

This enthusiasm is unique because it persists even though Alphabet's accelerated spending has pressured free cash flow in recent periods. This paradox resolves when investors consider the long-term economics: Heavy capital outlays today lay the foundation for proprietary AI services that should generate superior returns once the infrastructure is fully deployed.

Billionaires like Druckenmiller appear willing to tolerate near-term compression in cash flow because they see the alternative -- ceding technological leadership -- as more costly in the long-run. With this in mind, Druckenmiller's rotation away from Broadcom and recent investment in Alphabet looks less like a simple swap and more like a bet on ownership of a complete AI stack.

Adam Spatacco has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Berkshire Hathaway, and Broadcom. The Motley Fool has a disclosure policy.

As a former top lieutenant to George Soros, Stanley Druckenmiller has built a reputation for delivering big long-term returns through disciplined macro analysis. His primary investment vehicle today is the Duquesne Family Office, which manages a portfolio spanning technology, healthcare, and select cyclical themes.

Investors watch his moves closely because his track record and bold willingness to rotate positions decisively have often preceded major market shifts. The firm's latest 13F filing shows one such rotation: During the second quarter, Duquesne fully exited its position in Micron Technology (NASDAQ: MU) while initiating a new stake in Advanced Micro Devices (NASDAQ: AMD).

Stanley Druckenmiller. Image source: Getty Images.

Why sell Micron stock now?

Micron designs and manufactures advanced memory and storage solutions, including high bandwidth memory (HMB), DRAM, and NAND, that sit at the heart of artificial intelligence (AI) servers. Memory solutions feed data to graphics processing units (GPUs), keeping large language models (LLMs) running efficiently.

Throughout 2026, Micron stock has staged one of the market's most dramatic rallies -- rising 231% and achieving a trillion-dollar market capitalization. Such parabolic ascents often prompt seasoned money managers to take profits. Druckenmiller's complete exit from Micron may signal a view that memory is a more cyclical, capacity-driven segment of the broader AI chip stack.

While HBM demand is real, supply responses could arrive faster relative to next-generation specialized processors designed by AMD. In turn, this could potentially cap further valuation expansion for Micron if growth decelerates. By stepping away after the initial surge, Druckenmiller appears to be treating pure-play memory producers as a trade that is reaching maturity, rather than a multi-year compounder within the AI infrastructure landscape.

The case for buying AMD stock right now

A few years ago, Druckenmiller had built and subsequently fully sold a position in Nvidia. He later acknowledged that the exit was a "big mistake" as he left substantial gains on the table. I think this experience may partially explain the fresh interest in AMD, as the company is Nvidia's closest peer and still scaling the AI adoption curve.

AMD's Instinct accelerators and EPYC server processors are gaining respectable market share as hyperscalers diversify their capex budgets beyond Nvidia's processors. Nevertheless, AMD stock has not yet commanded the same dominant narrative that was once reserved for Nvidia.

During Q2, the company generated revenue of $11.5 billion, up 50% year over year. Meanwhile, AMD's data center segment more than doubled to $6.7 billion and accounted for nearly 60% of total sales. What's most encouraging is that management guided for continued acceleration into the second half of the year as AMD continues to onboard hyperscaler demand.

Should you follow Druckenmiller's lead and buy AMD stock?

AMD stock has already delivered an impressive performance this year, advancing more than 120%. At current levels, AMD trades at a forward price-to-earnings multiple (P/E) of around 63. To put that into context, the broader semiconductor industry boasts a forward P/E of around 26.

AMD PE Ratio (Forward) data by YCharts.

AMD's valuation clearly embeds optimistic expectations for aggressive AI-driven growth and profit margin expansion. Whether its premium is justified will depend on management's execution, measured by market share gains against Nvidia, successful product ramps, and capitalizing on the secular tailwinds supported by accelerating AI infrastructure spending.

For most investors, the prudent path is not simply mimicking Druckenmiller's decisions. Retail investors are best served by weighing AMD's strong fundamentals and competitive momentum against the risk that its elevated valuation leaves little margin for error.

I think building a modest position in AMD alongside a diversified basket of technology stocks, coupled with monitoring the company's quarterly progress, offers a balanced way to participate in the upside rather than simply copying one billionaire's latest filing.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Advanced Micro Devices, Micron Technology, and Nvidia. The Motley Fool has a disclosure policy.

Stanley Druckenmiller is one of the most closely followed investors in modern financial history. After his extraordinary run managing the Quantum Fund alongside George Soros, he now runs the Duquesne Family Office. Investors pay close attention to Druckenmiller's moves because he has a history of identifying major macroeconomic shifts early and sizing positions aggressively.

Yet even legends make blunders. Druckenmiller, for example, has openly called his decision to sell Nvidia (NASDAQ: NVDA) when he did a "big mistake," noting that the stock continued climbing long after he exited his position. This admission raises a similar question: Did he just make a similar error with his early exit from Micron Technology (NASDAQ: MU)?

Image source: Getty Images.

Breaking down Druckenmiller's Micron trade

Duquesne initiated a position in Micron during the first quarter of 2026, acquiring 23,400 shares. The firm's most recent 13F form, filed with the Securities and Exchange Commission last week, reveals that sometime in the second quarter, Druckenmiller completely closed that position. While the exact entry prices and trade dates remain private, Micron's broader price action is clear. During the first six months of the year, Micron stock gained more than 300%.

MU data by YCharts.

In that same stretch, the memory-chip maker entered the trillion-dollar club -- a milestone also achieved by peers SK Hynix and Samsung. The scale of this move turned Druckenmiller's modest stake in Micron into a multibagger in just a matter of months.

What may have prompted Druckenmiller to sell Micron stock

Micron's surge has been fueled by unprecedented demand for high bandwidth memory (HBM) and advanced DRAM solutions that feed data into the training and inference engines of artificial intelligence (AI) models. Memory sits at the center of the hyperscaler compute build-out -- without adequate data storage capacity and bandwidth, even the most sophisticated accelerators would frequently have to sit idle while awaiting information to process.

Micron has been booking record revenues and expanding its margins to sky-high levels, and the new multiyear supply agreements it is inking with its large clients appear to be rewriting the narrative in the memory market away from its traditional boom-and-bust cycle. Against this backdrop, Druckenmiller's decision to sell into strength might appear to have been premature.

Smart investors will realize that there are several practical considerations that support taking gains now, however. Micron and its rivals are racing to bring new fabrication lines online. Memory markets, however transformed, are still going to respond to capacity additions and changes in the supply-and-demand dynamic. This makes a valuation that expanded threefold in a single quarter vulnerable to significant mean-reversion risk sooner than some investors may anticipate. By locking in profits, Druckenmiller reduced his exposure to any near-term inventory digestion or competitive response in memory supply chains.

Was history repeating?

It's hard to dismiss the parallel between Druckenmiller's early exit from Nvidia and his decision to sell Micron. Both companies rode the AI wave, each delivered a triple-digit-percentage gain over a short period, and Druckenmiller exited both stocks while their momentum remained strong.

A more thorough comparison favors a different conclusion, though. Nvidia's competitive advantages are supported by its ecosystem -- a moat that has proven extraordinarily durable. The tight integration between the company's GPU architecture and its widely used CUDA software platform has locked developers into its ecosystem in a real way -- creating switching costs that pure-play memory producers can't replicate.

Memory remains fairly a commoditized product, even at the high end. This is one reason why historically, after every boom cycle, every player in the space eventually feels the same headwind when supply catches up to and exceeds demand. Druckenmiller's exit from Nvidia occurred after the stock had already multiplied and its valuation looked extreme by traditional metrics.

With Micron, the time frame around its valuation expansion was much shorter. Moreover, the structure of the memory industry differs enough from the advanced processor market to justify investing with caution. Taking profits after a parabolic advance does not automatically mean Druckenmiller made an error.

In the long run, history may prove that Druckenmiller sold Micron too early. For now, his decision rests on some distinguishable logic compared to Nvidia rather than an identical misjudgment. In an AI market defined by rapid narrative shifts, the greater risk might have been refusing to bank abnormally high gains when given the opportunity.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Micron Technology and Nvidia. The Motley Fool has a disclosure policy.

After 14 consecutive quarters as a net seller of equities, Berkshire Hathaway's (NYSE: BRKA) (NYSE: BRKB) cash reserves dropped from nearly $400 billion at the end of the first quarter to roughly $365 billion by June 30, marking a clear strategic pivot under new CEO Greg Abel.

During the second quarter, the investment conglomerate purchased about $23.5 billion in stocks while selling only $3.7 billion -- producing net buying activity of nearly $20 billion. These moves, combined with Berkshire's recent share repurchases and selective acquisitions, signal that the company's leadership finally sees some attractive opportunities after years of patience and cash accumulation.

Image source: The Motley Fool.

Breaking down Berkshire's portfolio

Earlier this year, Abel oversaw a significant cleanup of Berkshire's portfolio, trimming or exiting several smaller positions to concentrate capital in higher-conviction holdings. According to Berkshire's 13F filings, sales included substantial reductions in Bank of America, Capital One, Kroger, DaVita, Ally Financial, and Nucor, as well as complete exits from Constellation Brands and Amazon.

On the acquisition front, Berkshire closed its $9.7 billion purchase of Occidental Petroleum's chemicals business in January and completed the $6.8 billion all-cash acquisition of homebuilder Taylor Morrison last month.

Share buybacks are also ramping up, totaling more than $4 billion during the second quarter alone. Taken together, these actions reduced Berkshire's cash pile while reallocating capital into both wholly owned businesses and public companies.

BRK.B Stock Buybacks (Quarterly) data by YCharts

Alphabet emerges as a magnificent holding

The new commitment that stands out in Berkshire's portfolio is Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG). Berkshire first established a position in the internet giant during the third quarter of 2025 and has steadily increased its exposure in 2026.

During the second quarter, Berkshire dramatically expanded its stake in Alphabet. A pivotal piece was executing a $10 billion private placement in June, split evenly between Alphabet's Class A and Class C share classes. The company made additional open-market purchases to further enlarge the position. What began as a modest foothold has swiftly become a core holding in the portfolio, reflecting conviction in Alphabet's long-term competitive advantages and growth trajectory.

How to interpret Abel's measured signal

Berkshire's reduction in cash and return to net equity buying do not signal a broad market bottom or an abrupt change in the company's investment philosophy. Remember, Berkshire still holds more than $360 billion in liquidity -- preserving its fortress balance sheet.

The recent buying activity suggests Abel and his leadership team have identified a specific value that outweighs the safety of short-term Treasuries. Alphabet's elevation to core status, alongside incremental acquisitions and stock buybacks, underscores a long-standing preference for durable competitive moats and reasonable valuations.

Ultimately, Berkshire's cash deployment indicates that Abel is prepared to put capital to work when the right opportunities emerge, while remaining characteristically disciplined about price and long-term appreciation.

Bank of America is an advertising partner of Motley Fool Money. Ally is an advertising partner of Motley Fool Money. Adam Spatacco has positions in Alphabet and Amazon. The Motley Fool has positions in and recommends Alphabet, Amazon, and Berkshire Hathaway. The Motley Fool recommends Capital One Financial, Constellation Brands, Kroger, and Occidental Petroleum. The Motley Fool has a disclosure policy.

When executives decide to put millions of their personal dollars into shares of their own company, those moves tend to carry a weight that few other signals can match. In early August, Intel (NASDAQ: INTC) CEO Lip-Bu Tan purchased 105,263 shares of his company's stock at $95 per share.

In an artificial intelligence (AI) landscape hallmarked by rapid technological shifts and intense competition, this open-market purchase invites closer examination of both Tan's motives and the broader narrative surrounding Intel's turnaround.

Image source: Intel.

Why do insiders buy stock in their own company?

C-suite executives and members of corporate boards buy stock in their own companies for reasons both practical and symbolic. At a basic level, insider purchases align their personal financial interests with those of outside shareholders. Unlike stock-based compensation, an open-market buy requires an actual outlay of personal capital.

When the buyer happens to be the CEO, the signal is particularly potent because the person with the deepest visibility into a company's daily operations, product roadmaps, customer pipelines, and competition is choosing to increase their exposure at the current market price. Investors tend to interpret these transactions as an expression of genuine confidence rather than obligatory optics or marketing.

Against this backdrop, I think Tan's recent purchase functions as a public declaration that Intel's trajectory justifies a substantial personal bet.

The timing of Tan's buy is important

The timing of Tan's purchase makes it even more noteworthy. Intel reported its second-quarter 2026 results in late July. The company reported total revenue of $16.1 billion, representing 25% year-over-year growth -- marking the company's strongest quarterly growth in more than 15 years. Adjusted earnings per share (EPS) were $0.42 per share, double the consensus estimate.

The data center and AI segment surged 59% to $6.3 billion, while client computing and physical AI contributed $8.9 billion, up 13%. Management's guidance for the third quarter was for revenue in the $15.8 billion to $16.8 billion range, and non-GAAP (generally accepted accounting principles) EPS of $0.38.

Just a few weeks later, Tan stepped into the market to buy stock at $95 per share. I think the proximity between Intel's report and Tan's purchase was deliberate. After a strong print, the CEO's purchase could suggest that he views Intel's post-earnings stock price as still attractive relative to the multiyear opportunity he sees in AI-driven compute, foundry expansion, and manufacturing leadership.

Should you buy Intel stock right now?

Intel stock has delivered a dramatic rerating throughout 2026. Shares have climbed 184% so far this year, but recently retreated amid broader market volatility and on news of a $20 billion equity offering. When shares pulled back into the mid-$90 range, it created the window during which Tan bought the dip.

In this specific instance, the action carried significance because it coincided with the company's capital raise -- during which the same $95 price was offered to the public. By participating at this level, Tan is effectively endorsing the valuation the company used to raise new capital for additional manufacturing capacity.

All told, Tan appears confident that Intel's operational momentum from the second quarter will continue to compound, eventually justifying higher valuations. Nvidia's (NASDAQ: NVDA) $5 billion investment in Intel, announced in September as part of a broader strategic AI partnership, underscores external validation of the same recovery trajectory that Tan is now personally underwriting.

Whether that optimism proves correct will ultimately depend on the company's execution in the foundry business, sustained gains in data center CPU market share, and the broader hyperscaler capex cycle. For now, however, Intel's CEO has placed a clear, personal wager that the company's turnaround is still in its early innings.

Adam Spatacco has positions in Nvidia. The Motley Fool has positions in and recommends Intel and Nvidia. The Motley Fool has a disclosure policy.

Space Exploration Technologies (NASDAQ: SPCX) completed its long-awaited initial public offering (IPO) in June. SpaceX stock opened on the Nasdaq at $150, delivering an immediate 11% lift from the offering price of $135. The company ultimately raised about $85.7 billion, far above the previous record set by Saudi Aramco.

In the days that followed, shares of SpaceX continued climbing -- closing the first session near $161 and briefly touching an intraday high of more than $225 shortly thereafter. The initial enthusiasm, however, was fleeting. Subsequent trading sessions came with selling pressure as early momentum faded and valuation questions mounted.

As of this writing (Aug. 17), SpaceX stock has rebounded to its first-day trading range of about $150. This trajectory underscores a familiar pattern among high-profile IPOs: An opening-day pop driven by scarcity and narrative, followed by a period of price consolidation in which reality and expectations start to align.

Image source: Getty Images.

How do IPO stocks typically perform?

During the past 15 years or so, there have been a number of major technology IPOs that show a consistent pattern of significant drawdowns during the first year as a public company. Across a cohort of roughly 30 companies, the median maximum share price decline within the first 12 months of being public was 54%.

Facebook (now Meta Platforms) experienced a peak-to-trough decline of 54% in its first year as a public company. Uber fell as much as 68%. Airbnb held up better on a relative basis, yet still retreated 39% at its worst point. Even eventual long-term winners such as Shopify and Snowflake endured declines in the mid-50% range before recovering.

These figures are not outliers. Rather, they represent a typical experience for heavily hyped IPO stocks once the initial liquidity event ends and lock-up expirations, earnings scrutiny, and competitive realities take center stage.

Where is SpaceX stock headed over the next year?

Applying the same historical trends to SpaceX stock points to a wide range of plausible outcomes over the next year. Across the cohort mentioned above, the median one-year return from the IPO price stood at negative 9%, while the average return was 14%. The maximum declines, in contrast, represent a worst-case peak-to-trough scenario.

Under the median-return path, shares would finish their first year at about $135. In contrast, the average-return path would propel them closer to $170. In a more severe case that mirrors the historical maximum-drawdown experience, SpaceX stock could temporarily bottom at about $100, or even about $70 if measured from today's levels.

Taken together, these scenarios imply a realistic 12-month trading range between $100 and $170. Even at the lower end of this band, SpaceX would still boast a frothy enterprise value, since the business is still scaling revenue and hemorrhaging cash.

Is SpaceX a good investment right now?

It's important to understand that none of the historical analogies explored above are guaranteed. Markets can and often do reward exceptional execution, and SpaceX's vertical integration featuring regular rocket launches, Starlink growth, and artificial intelligence (AI) ambitions is unique.

In my eyes, a more probable near-term path for SpaceX stock will involve continued volatility as the first major lock-up agreements expire, letting insiders and early investors sell their shares, and the company works to prove that its AI and space-based compute initiatives justify a premium valuation.

Against this backdrop, I think smart investors are better served waiting for a more attractive entry point, rather than buying into momentum at current levels. SpaceX stock has already shown how quickly early gains can evaporate. Patience for a deeper consolidation offers both a lower valuation and a clearer view of whether the underlying business can convert its inspiring narrative into durable free cash flow.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Airbnb, Meta Platforms, Shopify, and Snowflake. The Motley Fool recommends Uber Technologies. The Motley Fool has a disclosure policy.

Ray Dalio has long stood as an influential sounding board for investors. As the founder of the world's largest hedge fund, Bridgewater Associates, Dalio's reputation is supported by his ability to dissect economic cycles with unusual clarity. His outlook often serves as a guide for both institutional and individual investors.

In a recent appearance on The Diary of a CEO podcast, Ray Dalio said that the current euphoria surrounding artificial intelligence (AI) has produced "classic signs" that a bubble is forming. While Dalio acknowledges that AI technology is transformative, he argues that some stock prices mirror the speculative excesses that preceded two of history's most severe market collapses in 1929 and 2000.

Ray Dalio. Image source: Getty Images.

Analyzing valuation extremes

Dalio's warning of a bubble can be supported when viewed through the lens of the cyclically adjusted price-to-earnings (CAPE) ratio. This metric smooths earnings over a 10-year time period and adjusts for inflation. Currently, the CAPE ratio sits near 41. This reading exceeds the 32.6 level reached just before the 1929 crash and is within shouting distance of the all-time high of 44.2 recorded at the height of the 2000 dot-com era.

S&P 500 Shiller CAPE Ratio data by YCharts.

As shown above, elevated CAPE readings have historically signaled muted future returns. In 1929, the belief that a permanent era of prosperity fueled rapid industrialization eventually collapsed when confidence cracked. This triggered bank failures and ultimately resulted in the Great Depression. Back in 2000, internet companies with little or no sales or profits commanded astronomical valuations based solely on exciting growth narratives. Once capital dried up and reality set in, the Nasdaq (NASDAQINDEX: ^IXIC) lost roughly three-quarters of its value.

^IXIC data by YCharts.

Why Dalio's parallel matters

Today's CAPE level of 41 places the S&P 500 (SNPINDEX: ^GSPC) in rarefied territory. The parallel is not simply numerical. Just as railroads were set to revolutionize travel and commerce and the internet promised endless digital disruption, AI is now being marketed as an unstoppable force that justifies any valuation.

Dalio points out that paper gains are ballooning far beyond the actual absolute dollars flowing through the financial system. These dynamics inherently create a fragility in which a sudden need for liquidity can force widespread selling. The risk here is not that new technology will fail to deliver productivity gains, but that many AI stocks priced for perfection will correct sharply if growth expectations are not met or when liquidity tightens.

What can investors learn from Ray Dalio?

The grey-shaded columns represent recessions in the chart below. While history shows that elevated valuations eventually compress, the S&P 500 has always trended higher over the long term.

^SPX data by YCharts.

Dalio is not saying that a stock market crash is guaranteed to arrive tomorrow. He is simply reminding investors that upward biases have never been immunized against multi-year drawdowns. The most prudent response is not one of panic selling, but rather disciplined preparation. Maintain diversification across industries, especially durable sectors such as consumer staples, and hold sufficient cash to meet financial responsibilities without forced stock sales.

Investors who treat the AI revolution as one more cycle rather than a generational exception will be better positioned, whether the market continues to climb or eventually revisits the painful corrections of 1929 and 2000.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.

Palantir Technologies (NASDAQ: PLTR) has emerged as one of the biggest darlings of the artificial intelligence (AI) revolution. Demand for the company's Artificial Intelligence Platform (AIP), which features Palantir's Foundry, Gotham, and Apollo software suites, is off the charts from both the public sector and private commercial enterprises.

Currently, Palantir trades at a price-to-sales (P/S) ratio of 74. This valuation comes amid the company's rapid expansion, with recent quarterly revenue growth exceeding 90% year over year. The question smart investors are asking is what has happened in the past when software-as-a-service (SaaS) stocks reached comparable multiples, even while generating similarly aggressive growth.

Image source: Getty Images.

Analyzing high-valuation software stocks

Several high-profile SaaS companies have experienced trajectories similar to Palantir's. Between 2020 and 2021, shares of data warehousing specialist Snowflake surged to $401. This translated into a peak P/S multiple of roughly 221 during the stock's ascent. Cloudflare commanded a similar P/S multiple above 100 times during its late-2021 high. Meanwhile, Datadog exhibited a peak P/S near 70 during this same time frame.

SNOW PS Ratio data by YCharts. PS Ratio = price-to-sales ratio.

While revenue continued to expand sharply for each of these SaaS leaders, their respective stock prices eventually normalized -- falling upwards of 70% from their peaks and remaining subdued for years. These outcomes demonstrate that extreme valuation expansion struggles to persist once growth expectations face friction or until a new catalyst emerges.

SNOW data by YCharts.

Why valuations tend to compress

It's important to acknowledge that the multiples witnessed throughout 2020 and 2021 stemmed directly from the pandemic. Remote-work environments fueled a surge in demand for collaboration software, cloud infrastructure, and digital productivity tools. These needs accelerated SaaS adoption beyond normal industry trends.

Yet even without these extraordinary tailwinds, each of the companies above continued to deliver impressive growth rates after peak pandemic-related concerns subsided. Nevertheless, none of these companies sustained their multiples. The mechanism is straightforward: An expanding P/S ratio assumes that revenue will compound at abnormally high rates for many years without interruption.

In reality, all businesses eventually encounter competition, saturating markets, or macroeconomic shifts. In turn, sales growth moderates toward more normalized levels. As a result, investors usually re-rate the stock downward.

The lesson here is to understand that growth rates do not immunize stock prices. Rather, they tend to delay the inevitable outcome until the market no longer prices in perfection. The examples above illustrate that once valuation multiples exceed comparable thresholds, subsequent returns often lag or turn negative, even while revenue and profits advance.

What does this mean for Palantir stock?

Palantir's current valuation profile mirrors the cases more closely than it diverges from them. Indeed, the company's commercial and government platforms are delivering exceptional growth, all while profit margins expand. Nevertheless, history suggests that Palantir's valuation assumes this trajectory will remain for an extended period. However, the precedents analyzed above prove that any deceleration, competitive response, or change in investor sentiment can swiftly trigger a rapid sell-off.

I think the actionable takeaway regarding an investment in Palantir can be found in the historical record above. At 74 times sales, Palantir may be positioned more for multiple compression than bulls realize. In turn, this could leave Palantir stock range-bound or even lower over the next couple of years, even if the company continues riding AI-driven tailwinds.

Investors with a concentrated position in Palantir may want to consider trimming exposure or reallocating to other software names with more moderate valuations. Meanwhile, long-term believers should prepare for a period of limited share price appreciation until sales catch up with the surging stock price.

Adam Spatacco has positions in Palantir Technologies. The Motley Fool has positions in and recommends Cloudflare, Datadog, Palantir Technologies, and Snowflake. The Motley Fool has a disclosure policy.

Alphabet's (NASDAQ: GOOG) (NASDAQ: GOOGL) long-running bet on Space Exploration Technologies (NASDAQ: SPCX) has quietly become one of the most consequential corporate investments in modern history. What began as a shared interest in satellite connectivity has now become a windfall that dominates Alphabet's financial profile. The analysis below details the power of deploying patient capital and the distortions that unrealized gains can introduce into reported profits.

Image source: Getty Images.

Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »

The origins of Alphabet's investment in SpaceX

In 2015, Google invested $900 million into SpaceX. At the time, the rocket company was valued at roughly $12 billion, so the investment secured Google an ownership stake of approximately 7.5%. The capital was used to support SpaceX's ambitions in reusable launch cadences and its nascent Starlink constellation. These areas aligned with Google's own interest in global internet access.

Over the last decade, Google's position was diluted through subsequent funding rounds. However, the company retained a meaningful stake in SpaceX. According to recent filings, Google's early check has now grown more than 100x in value, illustrating how a single investment can transform a balance sheet years later.

Breaking down SpaceX's landmark IPO

SpaceX completed an initial public offering (IPO) in June. According to its S-1 filing, SpaceX offered 555.6 million shares at a price of $135 each -- planning to raise $75 billion at a $1.8 trillion valuation.

In reality, SpaceX stock opened well above the offering price and closed its first day of trading near $161. This propelled the company's market capitalization past $2 trillion, instantly making it one of the most valuable companies in the world. On the last day of the second quarter (June 30), SpaceX shares were at $170.86.

How to assess Alphabet's Q2 earnings

For the quarter ended June 30, Alphabet reported net income of $112.2 billion on revenue of $119.8 billion. At first glance, this looks almost unbelievable. But a quick look at Alphabet's income statement reveals that the company's bottom-line expansion was almost entirely driven by a line item called other income, which totaled $98 billion.

Smart investors understand that companies often bury important notes and disclosures deep in their filings. According to Alphabet's latest 10Q, "other income" captures net gains on equity securities. Alphabet revealed that the surge from other income was "primarily related to unrealized gains in our equity securities portfolio from SpaceX and a private company." According to Alphabet's quarter-end 13F filing, the company's SpaceX position was worth $94.1 billion.

If I subtract SpaceX's equity gains, Alphabet's reported net income would move closer to $18 billion. This would actually have resulted in a 35% year-over-year decline in earnings per share (EPS). This distinction is important because unrealized gains are non-cash and vulnerable to daily stock price fluctuations. A subsequent decline in SpaceX stock -- which has since happened since the quarter ended -- essentially reverses the same line item that drove most of Alphabet's profitability in the first place.

Against this backdrop, investors should treat reported profits with an extra level of scrutiny, especially if meaningful equity positions are marked to market value. For Alphabet specifically, the most relevant metrics remain operating income, free cash flow, and the trajectory of its advertising and cloud computing segments.

While the SpaceX stake is a genuine economic asset, its contribution to quarterly financial results is inherently episodic and largely outside of the control of Alphabet's management. Smart investors should focus on the durable, cash-generating segments of Alphabet's ecosystem rather than the valuation swings of an investment portfolio. In the long run, this approach provides a clearer view of Alphabet's underlying health and earnings power.

Adam Spatacco has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet. The Motley Fool has a disclosure policy.

Alphabet's (NASDAQ: GOOGL) latest earnings report put two enormous sums front and center: a full-year capital expenditure guidance range that it increased to as much as $205 billion and a Google Cloud backlog that has climbed to $514 billion.

The scales of these figures invite comparison -- which one should investors weigh more heavily? The answer becomes more clear when these numbers are understood as two sides of the same coin.

Alphabet is pouring unprecedented sums into artificial intelligence (AI) infrastructure precisely because customer demand -- quantified by its towering backlog -- is accelerating. One number represents its investments, while the other is proof that the investments are paying off.

Image source: Alphabet.

Where is Alphabet's capex going?

Alphabet's AI infrastructure budget will be directed toward servers, GPUs, CPUs, memory, custom chips called Tensor Processing Units (TPUs), data center construction, and the networking gear that stitches everything together. Roughly 60% of the company's recent capital outlays went into servers, while the remaining 40% funded facilities and connectivity.

The importance of Alphabet's rising capex is straightforward. Without additional compute, the company will struggle to convert the capacity agreements it has already inked into revenue. In an environment where AI workloads are expanding faster than traditional cloud usage, underinvesting in AI development would cede ground to rivals -- namely Amazon Web Services (AWS) and Microsoft Azure.

Alphabet holds more than $240 billion in cash and marketable securities on its balance sheet, providing it with the financial flexibility to fund its AI build-out even while its free cash flow turns temporarily negative.

Understanding Google Cloud's backlog

Google Cloud's backlog did not pile up overnight. Rather, the half-trillion-dollar sum reflects a surge in multiyear enterprise commitments for AI-powered solutions. Alphabet CEO Sundar Pichai explained that roughly 90% of the Fortune 100 now use the company's Gemini Enterprise model in some form. He went on to explain that customer acquisition is doubling year over year as existing clients exceed their original consumption commitments by more than 50%.

Alphabet expects to recognize a little more than half of its current cloud backlog as revenue over the next 24 months. That schedule provides useful visibility to investors because it explains how a substantial portion of the infrastructure Alphabet is building today is effectively presold.

I expect sales from the company's TPU-based systems will ramp sharply going into 2027, while the remainder of the backlog will flow through ancillary Google Cloud Platform (GCP) services.

Breaking down Alphabet's virtuous cycle

When viewed in isolation, Alphabet's capex plan looks like an overzealous bet on an uncertain future. However, when viewed alongside the company's cloud backlog, it appears more validated, given an already visible future. Essentially, Alphabet's infrastructure budget covers buying servers and building data centers that will enable the company to meet pre-established capacity demand. In turn, Google Cloud generates both revenue and cash flow that justifies continued reinvestment in its AI ecosystem.

Revenue from Google Cloud accelerated 82% year over year in the second quarter, while the segment's operating margin expanded dramatically. This demonstrates that the early returns on prior AI infrastructure spending are materializing.

Ultimately, I think Alphabet's backlog is the more important figure for investors to focus on because it represents external validation that the company's internal spending is necessary. Spending on new programs alone does not create value. But smart capital allocation deployed toward durable, contracted AI-driven demand does.

Alphabet's AI story is not one of reckless spending or intangible growth. Rather, the company possesses a unique virtuous cycle in which AI infrastructure investments are translating into measurable, accelerating cloud adoption. As long as these dynamics hold up, I suspect both numbers will continue rising.

Adam Spatacco has positions in Alphabet, Amazon, and Microsoft. The Motley Fool has positions in and recommends Alphabet, Amazon, and Microsoft. The Motley Fool has a disclosure policy.

Have you ever wondered what would happen to your capital if you simply kept investing a modest sum into the S&P 500 (SNPINDEX: ^GSPC) every time the stock market had a meltdown?

Even without precisely timing the exact bottom of each crash, you'd still be looking at a pretty sweet pile of cash simply from investing during the thick of panic. That's the power of treating stock market slumps as opportunities rather than disasters. Let's walk through how this plays out and what smart investors can learn about buying the dip.

Image source: Getty Images.

Taking a look at each market crash since 1950

The stock market has experienced its share of rough patches during the past 75 years or so. For the sake of this analysis, I am specifically isolating events during which the S&P 500 dropped by 20% or more:

1. 1957: The index slid from the high 40s to 39.
2. 1962: The index fell from the low 70s to 52.
3. 1966: The index dropped from the mid-90s to 73.
4. 1970: The index decreased from more 100 to about 69.
5. 1973-1974: An oil shock caused the index to fall from about 120 to 62.
6. 1982: The index fell from the 140s to roughly 102.
7. 1987: During Black Monday and its aftermath, the index was chopped from the mid-300s to about 224.
8. 2000-2002: After the dot-com bubble burst, the index fell by nearly half from 1,527 to about 777.
9. 2008-2009: During the Great Recession, the S&P 500 plummeted from 1,565 to 677.
10. 2020: During the early days of the COVID-19 pandemic, the index fell from 3,386 to 2,237 over the course of a month.
11. 2022: The bear market of 2022, which featured historically high levels of inflation, dragged the index from 4,800 to roughly 3,577.

Running the numbers

In each case, I'm assuming that the series of $1,000 investments was not made at the absolute rock-bottom close. To be realistic, I'm treating each outlay as being roughly 10% higher than the trough. That would equate to the following entry points and gains:

Data source:

Here's the simple math behind the table. Each $1,000 investment divided by the entry price tells you how many "units" of the S&P you own. From there, I multiply those units by the current level of the S&P 500 -- about 7,750. In total, these 11 separate investments generated more than $750,000 in cumulative gains -- a return of more than 68-fold.

What is the lesson here?

Market stumbles always have and always will happen. But every single time, one common theme emerges: The market eventually rebounds, and the S&P goes on to new highs.

One of the most common aspects of behavioral finance is to panic sell when the outlook is bleak. But if you remain calm and disciplined, smart investors can use these resets as a chance to buy quality companies at a discount. From there, exercising patience is the key as the market's long-term ascent does the heavy lifting for you.

The big takeaway here is simple: Don't run for the hills when things get ugly. Instead, use dips to your advantage, and let time and compounding do the hard work. The S&P 500 has a long, resilient history of recovering and moving higher. Over the course of a long-term time horizon, the market ultimately rewards those who stick around.

Adam Spatacco has no position in any of the stocks mentioned. The Motley Fool has no position in any of the stocks mentioned. The Motley Fool has a disclosure policy.