Ahead of Lululemon's (NASDAQ: LULU) fiscal Q2 earnings report, I wrote an article published on Aug. 26 that said the stock looked like a value trap and that the warning from Dick's Sporting Goods would likely spill over and impact it as well. The stock subsequently plunged 17% on Sept. 4, in the session following its earnings report, as the athleisure company reported disappointing results and cut its full-year outlook. The stock has now lost more than half its value this year and nearly three-quarters of its value over the past five years.

Let's dive into the yoga brand's latest results and prospects to see what could come next for the once-high-flying apparel stock.

Troubles continue

Unfortunately for Lululemon, cutting guidance has become commonplace. For the fourth time since last June, it slashed its full-year outlook. It now expects revenue to decline by 7% to 5% to between $10.35 billion and $10.5 billion, down from prior expectations for sales in a range of $11 billion to $11.15 billion. Full-year adjusted EPS is projected to be between $9.48 and $9.73, but that includes a $0.86 tariff refund. Earlier, it guided to adjusted EPS of $10.95 to $11.15 without a tariff refund.

The company's Q2 results were pretty dreadful, and it looks like things are only worsening. Management noted everything from negative social media commentary to weak responses to new product launches to increased competition and brand deterioration.

Overall, the company's Q2 revenue fell 4% year over year to $2.42 billion, missing the $2.46 billion consensus estimate. Adjusted earnings per share (EPS) plunged 34% to $2.01, but were above the $1.79 consensus.

The underlying numbers were even worse. Americas revenue sank 8%, while same-store sales plunged 12%. International revenue rose 4%, but only 2% in constant currencies, while comparable sales in constant currencies slipped 6%.

China had long been a bright spot for Lululemon, but revenue fell 2% in constant currencies while same-store sales dropped 8% excluding foreign currency movements. The company said it was impacted by negative brand sentiment, which shouldn't be surprising given its big PR gaffe in China when, at an important yoga event held on the Great Wall, it inadvertently gave a Chinese actor a Japanese taiko drum to play instead of a Chinese dagu drum. Rest-of-world sales rose 6% in constant currencies, but comparable-store sales on the same basis dropped 6%.

Gross margin decreased by 200 basis points to 60.5%, but it would have been down 360 basis points when excluding the tariff refund.

Inventory was basically flat year over year, and it is doing a decent job of keeping this in check. This is an important metric to monitor for struggling brands, as big increases above sales growth can lead to more markdowns and sales.

Looking ahead, things will start getting worse for the company just as its new CEO takes over. While it is not uncommon to set a low bar when a new CEO or CFO comes on board, the company still projected a pretty meaningful sales decline. It expects Q3 revenue to decline by 10% to 11% to between $2.290 billion and $2.320 billion. Adjusted EPS is expected to fall to between $0.93 and $0.98 for the quarter, versus $2.59 a year ago.

Image source: The Motley Fool

Is the stock a buy on the dip?

While Lululemon stock looks cheap, now trading at a forward price-to-earnings (P/E) ratio of around 9 times this year's and next year's analyst estimates, the stock looks like it is set to fall into the same trap as other once very popular athletic apparel brands like Nike and Under Armour. The brand has lost its luster and faces increased competition, and, quite frankly, from my viewpoint, the athleisure fashion trend is shifting. I was recently eating lunch at Panera, and nearly everyone was wearing jeans. That is not something you would have seen a few years ago.

As such, this is a stock I'd still stay far away from, and it will likely take at least several years for a potential turnaround.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Nike. The Motley Fool recommends Lululemon Athletica Inc. and Under Armour. The Motley Fool has a disclosure policy.

After turning in another solid quarter, GitLab (NASDAQ: GTLB) is starting to prove the bear case wrong, and its stock is finally beginning to reflect that, with its shares climbing on its latest report.

The DevSecOps (development, security, and operations) company not only reported results that topped expectations and issued upbeat guidance, but its new annual recurring revenue (ARR) also grew at its fastest pace in several years. This is an indication that its core growth trajectory is reaccelerating.

Let's dig into the company's results and prospects to see why I think this growth stock remains a buy.

GitLab starts to flex its strength

GitLab turned in some impressive underlying metrics in the third quarter, led by its new ARR surging 42% year over year, its second-highest rate in the past four years. Its calculated billings rate, meanwhile, jumped 24%, which was double the growth rate it saw last quarter, and it said its sales team delivered its largest gross bookings ever. Its first-order count more than doubled to 1,700, while first-order ARR rose 39%. Meanwhile, its dollar-based net retention remained strong at 117% over the past 12 months, showing the first sequential improvement since 2024.

Long pegged as a loser amid the rise of artificial intelligence (AI), the company is starting to thrive in the current landscape. Management noted that AI is significantly lowering the bar for software development, which is helping drive demand for its platform and services. In addition, AI is giving GitLab more opportunities to monetize the growing amount of work occurring across the software life cycle.

The company recently introduced its Flex model, which lets customers commit to an annual dollar rate that it can shift between seats, consumption credits, and new capabilities. It expects this model to improve retention and drive growth, although it will have some revenue-recognition impact. It currently thinks that for every $50 million converted to Flex, it would lead to $5 million of revenue being recognized in future periods. Since its introduction six weeks ago, the company has already seen customers commit over $20 million to the program.

Turning to GitLab's results, overall revenue jumped 21% year over year to $286.3 million. That was well above the company's guidance for sales of $272 million to $274 million. Subscription revenue also increased by 21% year over year to $258.3 million, while license revenue rose by 20% to $27.9 million.

The company continues to see strength with its largest customers. Deals of $500,000 or more grew by more than 150% in the quarter. Sales of its high-end Ultimate tier, meanwhile, jumped 35% and now accounts for 59% of its ARR. It also said it saw a rebound in the public sector, which had been struggling.

Management once again upped its full-year guidance and now expects full-year fiscal 2027 revenue of $1.129 billion to $1.133 billion, representing growth of 18% to 19%, and adjusted earnings per share (EPS) in the range of $0.85 to $0.87. That's up from a prior forecast for revenue of $1.112 billion to $1.118 billion and adjusted EPS of $0.79 to $0.82.

For the fiscal third quarter, it forecasts revenue to be between $281 million and $283 million, representing 15% to 16% growth. It guided for adjusted EPS between $0.19 and $0.20. The company said it has not adjusted its guidance yet for the potential impact Flex could have on growth.

Image source: The Motley Fool.

The stock still looks like a buy

While off its lows, GitLab's valuation remains attractive. The stock is trading at a forward price-to-sales multiple of under 6.5 based on analyst estimates for fiscal 2028 (ending January 2028), despite the company growing its revenue around 20% and having over 15% of its market cap in cash.

Most importantly, the underlying metrics point to a business that is about to reaccelerate. While Flex will cause some distortions, that should not impact how investors view the stock. As such, I still consider it a buy even after its rebound.

Geoffrey Seiler has positions in GitLab. The Motley Fool recommends GitLab. The Motley Fool has a disclosure policy.

It wasn't too long ago that Snowflake (NYSE: SNOW) was viewed as a potential AI loser. Today, the company looks to be one of the biggest AI winners outside the infrastructure space. The stock recently surged 16.6% the session following its fiscal second-quarter earnings report and is now up nearly 70% on the year.

The cloud-based data warehousing and analytics company's architecture, which separates storage from compute to allow customers to store data and then process it across multiple cloud computing providers, is proving integral in the age of AI. Its solution has become an important system of record for agentic AI and also, importantly, allows for model choice.

Let's take a closer look at Snowflake's fiscal Q2 results to see whether the growth stock can keep its momentum or if it's too late to buy the rally.

Image source: The Motley Fool.

Snowflake's strong momentum continues

AI continues to be a big growth driver for Snowflake, with the company saying that it is at the center of the push toward enterprise agentic AI, as its platform "provides that trusted foundation." It's seeing rapid adoption of its AI coding agent CoCo and ready-to-use agentic app CoWork, while noting that its flexible model approach, which lets customers switch models and optimize costs, is a competitive advantage.

During the quarter, which ended July 31, the company's revenue climbed 35% year over year to $1.55 billion, topping the $1.48 billion analyst consensus. Product revenue, meanwhile, jumped 37% to $1.49 billion, its third-straight quarter of acceleration. Adjusted earnings per share (EPS) surged to $0.62 from $0.35 a year ago, easily surpassing the $0.45 consensus.

Snowflake continues to see strong expansion within its existing customer base, with net revenue retention rate coming in at 126% over the past 12 months, the same as in Q1. A number more than 100% indicates that existing customer usage is increasing after accounting for customer churn.

Snowflake also added 692 new customers in the quarter, including 14 Global 2000 companies. That was a 32% increase in net additions year over year. Meanwhile, it now has 828 customers who spend more than $1 million annually.

Snowflake raised its forecast for full-year product revenue to approximately $6.07 billion, up from previous guidance of $5.84 billion. The new outlook represents year-over-year growth of 36%. The company also raised guidance for its adjusted operating margin to 14.5% from 13.5%.

For fiscal Q3, it forecast product revenue between $1.588 billion and $1.593 billion, representing growth of 37% to 38%. It's looking for adjusted operating margin of 15.5%.

Is the stock still a buy?

Snowflake continues to fire on all cylinders. It continues to see great growth with existing customers, despite its large size, while it is also doing a great job of bringing on new customers.

It's truly positioned itself as an important model-agnostic platform that is paramount for the deployment of enterprise AI. With agentic AI still in the very early innings and the company continuing to build its own strong pipeline of AI products, Snowflake should have many years of strong growth in front of it.

The stock's valuation, though, is another story. With its strong performance this year, the stock now trades at a forward price-to-sales (P/S) multiple of 20 times this fiscal year's analyst estimates and 16 times fiscal 2028 (ending January 2028). That's toward the high end of its range since 2024, with similar to slightly higher revenue growth.

While I think Snowflake is positioned to be a long-term winner, its valuation could cap its near- to medium-term upside. As such, I wouldn't chase the stock here, but investors should be on the lookout to add shares on any meaningful pullback.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Snowflake. The Motley Fool has a disclosure policy.

With its artificial intelligence (AI) revenue surging and its stock recently stuck in the mud, Broadcom (NASDAQ: AVGO) is starting to look like one of the most attractive stocks in the AI infrastructure space. The designer, developer, manufacturer, and global supplier of semiconductor products is poised to see explosive growth in the coming years, with AI revenue projected to double in fiscal 2027 to $115 billion, up from earlier guidance of $100 billion, and then double again in fiscal 2028 to $230 billion.

Let's take a closer look at Broadcom's recent results and why the semiconductor stock looks like a great buy at current levels.

Image source: The Motley Fool.

Custom chips lead the way

Broadcom saw its AI semiconductor revenue skyrocket 221% year over year in fiscal Q3 ended Aug. 2. During the quarter, it delivered mass shipments of Ironwood TPU v7 chips to both Alphabet and Anthropic and began shipping Alphabet's next-generation Tensor Processing Units (TPUs) v8i to the company. Broadcom said it is handling Alphabet's TPU 8i version for inference, while MediaTek is dealing with the v8t version for training. Notably, it brought the 8i version to market more quickly than MediaTek, despite a later start.

The real story, though, was the company's outlook. The $230 billion AI revenue estimate for fiscal 2028 was well above the $180 billion estimate previously projected by Citigroup analysts. Meanwhile, it said its AI networking business would grow just as quickly as its custom chip business.

The company laid out where this revenue would come from, saying Anthropic would become its largest customer next fiscal year, with the frontier lab planning to deploy 5 gigawatts of TPU version 8i in fiscal 2027 and then 10 gigawatts in fiscal 2028. Alphabet will continue to be one of its most important customers, with it generating tens of billions of dollars in TPU revenue annually in the coming years.

OpenAI, meanwhile, is projected to become its second-largest chip customer, with 5 gigawatts of its new Jalapeño chip and its successor expected to be deployed in 2028. It will also deliver Meta Platforms 3 gigawatts of its custom MTIA chips through 2028, covering three generations.

Broadcom noted that its fiscal 2027 AI revenue could be higher, with demand currently above its $115 billion revenue projection, but that it needs to improve supply. For 2028, though, it has secured supply to meet its outlook. Given its huge AI revenue growth over the next two years, Broadcom now expects to generate adjusted EPS of over $30 in fiscal 2028.

Turning to the company's fiscal Q3 results, Broadcom's overall revenue surged 86% year over year to $29.6 billion, while adjusted earnings per share (EPS) soared 96% to $3.32. The results topped analyst expectations, with adjusted EPS of $3.24 and revenue of $29.36 billion, as compiled by LSEG.

Total semiconductor solutions revenue skyrocketed by 127% year over year to $20.8 billion. Its non-AI chip revenue growth remains sluggish, up just 5% to $4.2 billion in the quarter. Infrastructure software revenue, meanwhile, climbed by 29% to $8.8 billion.

Overall gross margin came in at 75%, down 210 basis points, as semiconductor revenue makes up a larger share of overall revenue. Software gross margin was 84%, compared with 76% for its semiconductor segment.

Looking ahead, Broadcom forecasts fiscal Q4 revenue to grow by 93% to $34.8 billion, with AI revenue surging 236% to $21.6 billion. Gross margin is expected to be 73%.

Time to buy the stock?

Broadcom is seeing surging revenue while, importantly, also diversifying its AI chip customer base. Despite that, the stock is now trading at a forward price-to-earnings (P/E) ratio of 11.5 times the fiscal 2028 guidance it just delivered. That's just incredibly cheap for what is becoming one of the best growth stocks in the AI infrastructure space.

As such, I'd be a buyer of the stock on this dip.

Citigroup is an advertising partner of Motley Fool Money. Geoffrey Seiler has positions in Alphabet, Broadcom, and Meta Platforms. The Motley Fool has positions in and recommends Alphabet, Broadcom, and Meta Platforms. The Motley Fool recommends London Stock Exchange Group Plc. The Motley Fool has a disclosure policy.

After sliding to start the year, Palo Alto Networks (NASDAQ: PANW) shares have come roaring back, trading up more than 80% so far in 2026. The cybersecurity stock benefits from the launch of Anthropic's Mythos models, which exposed previously unknown software vulnerabilities. The company said this has led to a shift in the security landscape and that it is just the start.

However, despite a strong recent earnings report and guidance, the stock price fell, as expectations were sky-high following its run-up this year. Let's dig into the company's latest earnings report and prospects to see if the stock's still a buy on this dip.

Image source: Getty Images.

Platformization strategy continues to pay off

While it was a bold move at the time, Palo Alto's platformization strategy, where it decided to forgo selling point solutions in favor of offering three cybersecurity platforms, continues to pay dividends. The company saw 220 net new platformization additions in fiscal Q4, double the 110 it saw in the prior quarter. Meanwhile, net revenue retention among these customers surpassed 120%.

Palo Alto said that Mythos has driven platformization demand, as customers are increasingly looking for a unified platform to tackle potential AI threats. Right now, there is a big push among organizations to have real-time defense, which it believes can only be achieved with a unified platform. As such, it sees AI as a significant growth tailwind for both itself and the broader cybersecurity industry.

During the year, the company also bolstered its platform through two large acquisitions to enhance its cybersecurity capabilities. First, it bought real-time data monitoring company Chronosphere, which it closed in January, and then it acquired privileged access company CyberArk. It said both are exceeding early expectations.

The combination of platformization and acquisitions helped drive strong growth for Palo Alto in its fiscal 2026 Q4, ended July 31. Revenue climbed 34% year over year to $3.41 billion, which was above the high end of its previous forecast for revenue of between $3.345 billion and $3.355 billion. Subscription and support revenue jumped by 36% to $2.67 billion, while product revenue rose by 29% to $738 million.

Next-generation security once again fueled Palo Alto's growth, with next-generation security annual recurring revenue (ARR) surging 63% to $9.1 billion. Network and AI security ARR rose 17% to $2.3 billion. Its Cortex Platform ARR rose 25% to $1.9 billion, with XSIAM (extended security intelligence and automation management) ARR surging 70%. Meanwhile, its Idira platform, which consists of its identity security platform from the CyberArk acquisition, contributed ARR of $644 million and was up 21% to $1.26 billion on an adjusted basis.

Adjusted earnings per share (EPS) increased by 7% year over year to $1.02, which was ahead of its guidance of $0.96 to $0.98.

Palo Alto forecasts fiscal 2027 Q1 adjusted EPS of between $0.96 and 0.98, with revenue rising 33% to 34% to between $3.3 billion and $3.31 billion. For the full fiscal year, it sees adjusted EPS coming in between $4.16 and $4.19 on a 23% to 24% climb in revenue to between $14.1 billion and $14.2 billion. It sees its next-gen security ARR rising 22% to 23% to a range of $11.075 billion to $11.175 billion.

Is Palo Alto stock a buy?

Even after the drop in its stock price, Palo Alto stock still trades at a hefty forward price-to-sales ratio (P/S) of 19.5 times fiscal 2027 estimates and a forward price-to-earnings ratio (P/E) of 81 times 2027 estimates. While the company is well-positioned and has some nice tailwinds behind it, that's a hefty price to pay for a company growing its ARR in the low- to mid-20% range, with acquisitions.

As such, I would not be a buyer on this recent dip.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool recommends Palo Alto Networks. The Motley Fool has a disclosure policy.

Artificial intelligence (AI) infrastructure stocks are still one of the biggest driving forces in the market. While Nvidia (NASDAQ: NVDA) helped lead the early charge, other stocks, like Micron (NASDAQ: MU), have greatly outperformed the AI leader over the past year.

Let's dive into both AI stocks to see which one looks set to outperform over the next five years.

Nvidia

Nvidia was the first big AI winner, and the advantages the company established remain in place today. Meanwhile, despite its massive size, the company continues to put up incredible growth. In Q2 2027, it more than doubled its revenue to a whopping $96.2 billion, with 117% growth in its data center segment. Meanwhile, it projected growth of 89% next quarter and 70% for fiscal 2028 (ending January 2028), while saying it remains capacity-constrained. In other words, it would be growing even faster if it could make more chips.

At the heart of Nvidia's growth story remains its graphics processing units (GPUs) and its CUDA software platform. Most early AI code was written using CUDA and optimized for its GPUs, which has given the company a wide moat in AI model training. That moat remains intact today, but the company is much more than just GPUs.

Today, Nvidia is a complete AI infrastructure company, offering full server racks designed for specific AI tasks, such as training, inference, and agentic AI. The company took a big leap in the inference market following its "acquisition" of Groq and its language processing units (LPUs). These chips, with a small amount of SRAM (static random-access memory) built in, can access memory quickly and reduce latency, making them great for the decode phase of inference, while its GPUs handle the more compute-heavy pre-fill phase. Meanwhile, it has also developed its own Arm-based central processing units (CPUs) specifically aimed at agentic AI. It then ties it all together with its powerful networking portfolio.

With Nvidia firmly established as the AI model training leader and well positioned for the fast-growing inference and agentic AI markets, this is a company that has a lot of growth in front of it over the next five years. Meanwhile, the stock is still attractively valued, trading at a forward price-to-earnings (P/E) ratio of 14 times analyst estimates for fiscal 2028 (ending January 2028).

Image source: The Motley Fool.

Micron

Micron has been riding the memory supercycle to astounding growth. The company saw its revenue last quarter surge to $41.5 billion from $9.3 billion a year ago, while its gross margin expanded to 84.6% from 37.7%. That led its adjusted earnings per share (EPS) to skyrocket from just $1.91 a year earlier to $25.11.

The revenue growth and margin expansion are coming from a huge supply-demand imbalance within the memory market. The company is one of the big three DRAM makers -- along with Korean companies SK Hynix and Samsung -- and derives around three-quarters of its revenue from DRAM and a quarter from NAND (flash). AI is driving memory demand, as GPUs are packaged with high bandwidth memory (HBM), a special form of DRAM that helps reduce latency and optimize performance, and AI training data is being stored in enormous solid state drives (SSDs) that use flash memory.

Rival SK Hynix has predicted that the DRAM market will remain supply-constrained through at least 2030, while Micron and its rivals have begun to lock in long-term contracts for up to five years for the first time ever. This should help reduce some of the extreme cyclicality the industry has seen in the past.

On a forward P/E basis, Micron looks very cheap, trading at only 6 times analyst estimates for fiscal 2027 (ending August 2027). The question is how long the memory cycle will last and whether there has been enough of a structural shift in the market to keep prices high.

The verdict

I think the easy answer here is that Nvidia will outperform Micron over the next five years. Micron is still a cyclical stock until proven otherwise, and five years is a long time for a cycle to last. Meanwhile, Nvidia has built one of the most complete and powerful AI infrastructure companies on the planet, and the stock is very attractively valued given its growth and moat.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Arm Holdings, Micron Technology, and Nvidia. The Motley Fool has a disclosure policy.

With about 40% short interest, SoundHound AI (NASDAQ: SOUN) is one of the most shorted stocks in the market. The company has been growing its revenue quickly, but a pending merger with troubled LivePerson (NASDAQ: LPSN) has investors betting against the stock. However, heavy short interest is a double-edged sword.

If a stock has high short interest, there is usually a bearish case to be made against it. In the case of SoundHound AI, its merger with LivePerson will bring considerable debt and a business currently in decline. There is also a good chance that SoundHound AI will seek a future equity raise to pay off the debt it is taking on as part of the deal, as it has stated it will work to quickly retire the debt through a combination of cash and stock.

The flip side is that heavily shorted stocks can see huge short squeezes if they can prove the skeptics wrong. With LivePerson, SoundHound AI is gaining access to a large, established, albeit declining, customer base, especially in call centers and the customer service space.

Its goal will be to retain these customers and switch them to its more comprehensive and higher-priced AI voice technology and agentic AI platform. If the company can stabilize churn and upsell customers, this deal could be a huge boon for this growth stock.

Image source: The Motley Fool.

SoundHound AI's core business, meanwhile, continues to perform well, and it's already shown it can integrate acquisitions well. Its earlier purchase of Amelia and its virtual agents helped it expand into new market verticals and provided key technology for its new voice-native agentic AI orchestration platform, OASYS.

The company has said that the launch of OASYS has increased excitement and accelerated deals, and that it is winning in demos, RFPs (requests for proposals), and pilots. The platform's ability to build AI agents and self-improve has been a big selling point, and it has been seeing pilots convert to large implementations at a record pace. Management noted that the platform has seen rapid adoption across verticals since its May launch, including a large eight-figure deal signed in less than 90 days after the initial demo.

Overall for Q2, SoundHound AI saw its revenue climb 45% to $61.9 million. However, the company does remain unprofitable, with an adjusted loss of $0.02 and negative adjusted earnings before interest, taxes, depreciation, and amortization (EBITDA) of $9.6 million. Its gross margin rose 610 basis points year over year to 45.1%, while its adjusted gross margins were flat year over year at 58.4%. It said acquisitions continued to have a temporary impact on its gross margins and that its goal is to still exceed 70% in the future.

Looking ahead, the company raised the low end of its full-year revenue guidance to $230 million to $260 million. That's up from a prior outlook of $225 million to $260 million. It is expecting its acquisition of LivePerson to close by the end of the year. If it does, it is projected to generate at least $350 million of revenue in 2027.

Is the stock a buy or sell?

SoundHound AI's stock has struggled over the past year, with its shares cut in half. However, the stock is still not cheap, trading at a forward price-to-sales (P/S) ratio of over 12 times current-year estimates. With the company not profitable and burning cash, that alone makes it a more speculative bet. The acquisition of LivePerson, meanwhile, adds some additional risk.

That said, the company has shown it can do a good job integrating acquisitions, and it has a huge opportunity with OASYS, both with existing customers and the ones it is acquiring from LivePerson.

Despite the huge short interest, I think investors can place a small speculative bet that SoundHound AI can pull this deal off, given the early momentum it has been seeing with OASYS.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends SoundHound AI. The Motley Fool has a disclosure policy.

Hyperscalers are companies that own massive data centers. While they are often associated with cloud computing providers, that isn't always the case, as some megacap tech companies still prefer to build out their own infrastructure to save costs.

Ultimately, these are the companies that are driving the AI infrastructure boom. They are spending massive amounts on capital expenditures (capex), and while pick-and-shovel plays like semiconductor companies are reaping the rewards, it is the deep-pocketed hyperscalers that are in control. After all, if they don't see strong returns on their AI infrastructure investments, they can cut off the spending spigot.

I currently own shares in three top hyperscalers -- Amazon (NASDAQ: AMZN), Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG), and Meta Platforms (NASDAQ: META) -- and think they are three of the best growth stocks to own over the long haul. All three have great core businesses with wide moats that generate massive operating cash flow that can help pay for their AI infrastructure spending, on which they are getting great returns. Here's why I own these three stocks.

Amazon

If you're like me, you're likely a frequent customer of Amazon. The company has built the nation's dominant e-commerce platform, and its warehouse and logistics network creates a wide moat that is unlikely to be challenged in the U.S. Meanwhile, its internal investments in robots and AI have created enormous operating leverage in the business.

What people may not realize, though, is that the business that contributes the most to Amazon's bottom line is actually its cloud computing unit, Amazon Web Services. The company created the entire infrastructure-as-a-service concept, and it remains the world's largest cloud provider today. The company also has some strong built-in advantages due to its custom chips, which help reduce inference costs, and its partnerships with leading frontier labs Anthropic and OpenAI.

Amazon is starting to fire on all cylinders again and has a big opportunity in front of it with a goal to eventually reach $1 trillion in cloud revenue.

Alphabet

Alphabet is best known for its Google search business, which continues to be a money-making machine. Meanwhile, the incorporation of its Gemini large language model into its platform and AI tools such as AI Overviews, AI Mode, Circle to Search, and Lens have been helping drive query and revenue growth. The company has one of the largest digital ad platforms on the planet, which is helping it drive consumer AI revenue better than other frontier labs such as OpenAI.

At the same time, its Google Cloud business has been growing at a rapid pace, with segment revenue surging 82% last quarter to $24.8 billion and operating income more than tripling to $8.8 billion. The company's Tensor Processing Units (TPUs), which are widely considered to be the best custom AI accelerators in the market today, give it a big cost advantage in training its own models, running inference, and offering a cheaper alternative to its cloud computing customers. It has even begun selling some TPUs directly to customers for deployments outside of Google Cloud.

Alphabet is the most complete AI company, and that positions it to be a long-term winner in the space.

Image source: Getty Images.

Meta Platforms

Meta Platforms is an unusual hyperscaler in that it currently does not operate a cloud computing business -- it built out its massive data center infrastructure solely to support its own compute needs. However, with demand for compute power so high, it is considering getting into the cloud business.

Few companies have been as good at using AI to drive growth in their core businesses as Meta, but it has a flywheel business built for AI. The company uses AI to improve its content recommendation engine to keep users on its platforms longer. This allows it to serve people more ads. At the same time, it is also giving advertisers better tools to connect with and convert users, which is helping it increase the prices it charges for ads. The better its models become, the more revenue it generates. In fact, it is believed that Meta could surpass Google this year to become the largest digital advertising platform in the world.

The stock has fallen to cheap levels over the past year or so due to market worries about the company's high AI infrastructure spending, and I think now is a great time to scoop it up. In fact, while I like all three of these stocks, Meta Platforms may have the most medium-term upside.

Geoffrey Seiler has positions in Alphabet, Amazon, and Meta Platforms. The Motley Fool has positions in and recommends Alphabet, Amazon, and Meta Platforms. The Motley Fool has a disclosure policy.

Artificial intelligence (AI) stocks have been leading the market higher over the past year, and some have gone absolutely parabolic with gains of 500% or more. However, that doesn't mean they may not have more upside ahead.

Let's look at three AI stocks with strong momentum.

1. Sandisk

Up a whopping 2,900% over the past year as of the end of August, Sandisk (NASDAQ: SNDK) shares have been on absolute fire the past 12 months. The company is a pure-play maker of NAND (flash) memory, and it has benefited from a wide supply-demand imbalance that has pushed NAND prices significantly higher. This, in turn, has led Sandisk's revenue to surge and its gross margin to balloon over the past year.

The current supply-demand imbalance stems from a few factors. First, after the NAND market crashed following the pandemic, when there was a lot of pull-forward in demand from electronics and computers, the big three memory makers slashed NAND capacity and directed it more toward DRAM (dynamic random access memory). Since then, these companies have mostly been largely focused on high bandwidth memory (HBM), which gets packaged with AI chips, like graphics processing units (GPUs), to optimize their performance. Soon after NAND capacity was cut, however, demand for massive solid-state drives (SSDs) that use flash memory started to take off to store AI training data.

While historically a highly cyclical business, Sandisk has started to sign long-term agreements. Its first three are worth a minimum of $42 billion, and it says it has five agreements in place for periods of up to five years. The deals cover one-third of its expected fiscal 2027 capacity, and it is looking to push that to over 50%. Meanwhile, the company projected that it could grow revenue in the mid-to-high teens between fiscal 2028 and fiscal 2030 while keeping its adjusted gross margin around 80%.

Despite its huge surge in price, the stock trades at a forward price-to-earnings (P/E) ratio of just 7. If it can sustain solid growth through 2030, the stock could have further upside.

Image source: Getty Images.

2. Micron

Up nearly 700% over the past year as of the end of August, Micron (NASDAQ: MU) is another memory maker that has seen its stock skyrocket. Micron is one of the big three memory makers, with 76% of its revenue last quarter coming from DRAM and 24% from NAND. The company is seeing similar dynamics as Sandisk, with memory supply-demand imbalances driving revenue growth and gross margin expansion.

The DRAM market is largely being driven by demand for HBM, which is growing in lockstep with demand for AI chips. However, supply remains constrained and looks like it will stay that way in the coming years. One reason for this is that the critical layers of HBM are manufactured using EUV (extreme ultraviolet lithography) machines, which are also used to make logic chips like GPUs. Meanwhile, ASML is the only company in the world that makes these machines, so there is a limit to the number it can make each year. At the same time, HBM requires upward of 3 times the wafer capacity of regular DRAM, which also limits growth and takes away capacity from ordinary DRAM. This has sent all DRAM prices skyrocketing.

Like Sandisk, Micron stock is also cheap, trading at a forward P/E of 6, and it, too, has locked in longer-term contracts. If this supercycle lasts into 2030 and beyond, the stock could have plenty of upside from here.

3. Lumentum

Up 550% over the past year, Lumentum Holdings (NASDAQ: LITE) is benefiting from the shift in AI data centers from copper wire to optical networks. It produces high-power indium phosphide (InP) lasers that convert electricity into light for high-speed data transmission. It is one of the few companies in the world capable of making these lasers in mass quantities, and holds up to a 60% market share for these advanced lasers. The company also has strong positions in the optical circuit switches (OCS) and co-packaged optics (CPO) markets.

Designing and manufacturing InP lasers is extremely difficult and requires specialized semiconductor fabs, as well as overcoming other challenges. Lumentum holds over 1,000 patents, and has decades of manufacturing scale that give it a yield and cost advantage. Meanwhile, once a laser or transceiver component is certified for an AI platform deployment, the risk of it being swapped out for a competitor is low.

With AI data centers just starting to move from copper wire to optical networks, Lumentum is in a strong position and could have nice upside from here.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends ASML, Lumentum, and Micron Technology. The Motley Fool has a disclosure policy.

Memory prices have been exploding higher over the past year, and with it so have the stocks of DRAM (dynamic random access memory) makers Micron Technology (NASDAQ: MU) and SK Hynix (NASDAQ: SKHY). However, the memory market is notoriously cyclical, and many investors are already questioning how long the good times will last for these AI stocks.

The typical memory cycle tracks a pattern of booming prices, followed by customers over-ordering, manufacturers increasing capacity, and then prices collapsing. The up cycle generally lasts a year or two, with DRAM makers seeing surging revenue and ballooning gross margins, followed by steep revenue and margin declines as new capacity floods the market.

Image source: The Motley Fool.

However, this does not look like your ordinary DRAM cycle. Past DRAM supercycles have historically been driven by surging demand for personal devices that use DRAM, such as smartphones and computers. The current DRAM supercycle is directly linked to the AI build-out. Unlike demand for personal devices, which cools off as consumer adoption matures, AI memory demand is being fueled by an arms race where tech giants must continuously expand compute capacity just to stay competitive.

The reason for this is that graphics processing units (GPUs) and other AI chips need to be packaged with high bandwidth memory (HBM), a special form of DRAM, to reduce latency and optimize performance. As such, demand for HBM is moving in lockstep with demand for overall AI computing power, which continues to surge. With the big three DRAM makers, which also include conglomerate Samsung Electronics in addition to SK Hynix and Micron, all focused on increasing HBM capacity, the overall DRAM market remains undersupplied.

At the same time, there are a few obstacles keeping DRAM makers from quickly ramping up capacity. First, the critical components of HBM and advanced logic chips, like GPUs, both use extreme ultraviolet lithography (EUV) in the manufacturing process, and there is only one company in the world, ASML Holding, that makes these machines. With foundries and DRAM makers both needing EUV machines, there is a limit to how much DRAM capacity can increase.

On top of that, HBM requires upwards of three times the wafer capacity as ordinary DRAM, which slows down capacity additions, as well. Clean room space also takes years to be built out, and foundries also need to build out more advanced packaging capabilities.

Peak earnings?

Right now, analysts have Micron's earnings peaking at $170.70 per share in fiscal 2028 ending August 2028, before falling to $121.77 per share in fiscal 2029 and then settling around a normalized level of around $50 per share. SK Hynix earnings are expected to hit $42.21 per share in 2028, and there aren't any outer-year estimates for its ADRs. However, SK Hynix management has said it does not see DRAM supply catching up to demand until at least 2030 at the earliest.

Now, another wrinkle to the market is that the big memory makers have all started to sign long-term, multi-year contracts for the first time ever. SK Hynix recently inked contracts worth $750 billion to supply customers with memory chips, including $500 billion from Nvidia. It and Nvidia will also work to co-develop next-generation memory linked to the chip giant's AI infrastructure roadmap.

Overall, the big three DRAM makers are all now looking to sign three- to five-year deals. Micron has said it has signed non-cancellable, long-term deals with 16 customers running through 2030 with defined price ranges that have a floor and ceiling. These contracts represent about 40% of its revenue. SK Hynix, meanwhile, reportedly does not have a price cap with its agreements.

How well Micron and SK Hynix perform in the coming years will largely be based on the length of the memory cycle and how soft the landing is. The new long-term contracts and source of demand from AI should make this cycle different from any in the past, as it is coming with a structural shift, but it is still likely a cycle.

In my view, the DRAM cycle is likely to extend a little longer than most analysts are anticipating, which should bode well for the stocks. Meanwhile, given its HBM leadership and ties to Nvidia, I think SK Hynix looks like the best memory stock to own over the long run.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends ASML, Micron Technology, and Nvidia. The Motley Fool has a disclosure policy.

While Greg Abel has taken the reins from Warren Buffett at Berkshire Hathaway (NYSE: BRKA) (NYSE: BRKB), the Oracle of Omaha's fingerprints are still all over the company's top holdings. Berkshire has held most of its top holdings for an extended period, while Buffett has explicitly said that he had a hand in the conglomerate's new top-three holding, Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG).

Buffett has long stated that he likes compounding businesses with wide, durable moats. With Berkshire's top three holdings making up more than 50% of its portfolio, let's dive into what Buffett most likes about these stocks.

Apple

Representing around 22% of its equity holdings, Apple (NASDAQ: AAPL) is Berkshire's largest position, and it's easy to see why Apple is a Buffett favorite. Quite simply, the company has one of the best compounding businesses on the planet.

The company has established itself as a high-end brand for smartphones, PCs, and other devices, all of which tend to have predictable replacement cycles. It also gives the company a more affluent customer base. Meanwhile, once consumers buy an Apple product, they generally become locked into the brand's ecosystem. Apple then generates high-gross-margin service revenue from things such as selling cloud storage, commissions on apps, Apple Pay, and a search revenue-sharing deal it has with Alphabet.

This all makes Apple a great compounding business with a strong, durable moat that is unmatched.

Image source: The Motley Fool.

American Express

American Express (NYSE: AXP), which accounts for over 17% of Berkshire's equity holdings, is another great compounding business that caters to affluent customers.

The business was built on its reputation for security and prestige. Before ever issuing charge cards, the company made a name for itself in the world of financial security as a freight delivery company moving highly valuable assets, and later it invented traveler's checks. When it finally began issuing charge cards, it positioned itself as a select club where members got certain privileges. This lets it charge its customers higher annual fees in exchange for these perks.

Unlike payment networks Visa and Mastercard, it operates a closed-loop network where it is both the card issuer and processor, which lets it capture revenue from both sides of the transaction. Because it caters to high-income consumers who spend three times the average of other consumers, it can charge merchants higher processing fees.

Meanwhile, the company tends to face low credit risk due to its affluent customer base and because a large percentage of its business is credit cards, where balances need to be paid off every month or incur interest charges. It's a great business that would be difficult to replicate today.

Alphabet

Berkshire's newest top-three holding is Alphabet, which accounts for nearly 13% of its equity holdings. This is another great compounding business with distinct built-in advantages, within both its search/AI discovery and cloud computing businesses.

Google remains Alphabet's largest business, and on that front, the company is seeing growth driven by new AI tools like AI Overviews and AI Mode. The company has a wide moat in search/AI discovery by controlling the distribution channels. This includes its ownership of the market-leading web browser (Chrome) and smartphone operating system (Android), as well as its revenue-sharing deal with Apple that makes Google the default search engine on Apple devices.

Alphabet also has one of the world's leading ad platforms, which helps it monetize consumer AI more effectively than most competitors.

Alphabet's fastest-growing business, meanwhile, is cloud computing. The company has a big edge here through its tensor processing units (TPUs), which are custom chips it designed over a decade ago and has built its entire software and hardware stacks around. This gives it a nice cost advantage over companies that are mainly reliant on Nvidia graphics processing units (GPUs). This also lets it train its models and run inference more cheaply. These models are used across Google Search and its other products, creating a nice flywheel effect.

As the most complete AI company, Alphabet continues to look well positioned for the future.

American Express is an advertising partner of Motley Fool Money. Geoffrey Seiler has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, American Express, Apple, Berkshire Hathaway, Mastercard, Nvidia, and Visa. The Motley Fool has a disclosure policy.

The market is set to see two of the largest IPOs ever in 2026. Space Exploration Technologies (NASDAQ: SPCX) debuted in early June with a nearly $1.8 trillion market cap and raised over $85 billion in capital. Anthropic, meanwhile, plans to IPO this fall, looking to raise $100 billion at a $2 trillion market cap.

But which of their stocks will be the better one to own?

SpaceX

SpaceX has been on a wild ride since its IPO, and the stock now sits about 5% above its $135 initial public offering price and about 4% below its first day opening price of $150. The stock is largely a bet on Elon Musk and the future, given the company's ambitious goals. At the time of its IPO, management pegged its total addressable market at a whopping $28.5 trillion, and Musk claimed the company would become worth more than Earth itself.

At the moment, the company's largest and most profitable business is its satellite internet service, Starlink. This business is growing quickly, adding subscribers and seeing strong revenue growth, including 66% in Q2 to $4.3 billion, while producing adjusted earnings before interest, taxes, depreciation, and amortization (EBITDA) of $2.6 billion.

SpaceX has talked about taking on traditional wireless carriers with Starlink, but its limited spectrum, the technical challenges of using satellite broadband in densely populated areas, and the time and cost required to build a nationwide terrestrial network would likely mean it would need to acquire a mobile provider for this to happen.

The company's space segment is its smallest by revenue, but perhaps its most important. The company has demonstrated it can launch its reusable Falcon rockets regularly, but the key will be to do so with its massive new Starship rocket. If it can recover and quickly reuse these rockets, it would greatly improve the economics of putting payloads into orbit and open a world of possibilities for the company's space ambitions.

AI is ultimately the company's biggest opportunity, and this includes data centers in space. There are still technical challenges to overcome, but the company is currently working with Nvidia to address some of them. Meanwhile, SpaceX is seeing great cloud computing economics as it leases its compute, often at high prices to hyperscalers that need the extra capacity.

Image source: The Motley Fool.

Anthropic

After a period of operating in the shadow of rival OpenAI, Anthropic burst into the spotlight as a leading model maker at the frontier. Instead of focusing on the consumer market, the company positioned itself as the leader in the enterprise and developer market with its long-term reasoning and coding agents.

Anthropic is particularly strong in developer-focused tool sets and agentic coding tools. This has enabled it to capture significant market share among enterprise customers that integrate its models via APIs. Its models are built on constitutional AI, resulting in more predictable AI behavior that appeals to regulated industries such as finance, healthcare, and law. Meanwhile, it uses a pay-per-token API pricing model to align its revenue closely with compute costs, enabling it to scale profitably.

The company has seen tremendous growth over the past year. In Q2, its revenue skyrocketed 14-fold year over year to $11.5 billion. Meanwhile, its annualized revenue run rate surged sevenfold to $65 billion at the end of July. It also turned in an operating profit, generating operating income of $559 million in the quarter. The company is really starting to separate itself from OpenAI, which generated $6.7 billion in Q2 revenue and had a $12.3 billion operating loss.

According to reports, Anthropic plans to cite a $30 trillion market opportunity in its prospectus, which seems like a number picked just to surpass the one given by SpaceX.

The verdict

SpaceX already carries a pretty hefty valuation, and while Anthropic hasn't gone public yet, given the numbers being thrown around, it likely will too. Meanwhile, both will have market caps that place them among the largest companies in the world. As such, I'd view both as speculative investments despite their massive sizes.

If I had to pick one of these two stocks to add to my portfolio, I'd go with Anthropic, as it looks well ahead in the enterprise AI race. Meanwhile, the fact that it has achieved hyperbolic revenue growth while remaining profitable is impressive.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Nvidia. The Motley Fool has a disclosure policy.

Artificial intelligence (AI) stocks have helped lead the market higher over the past few years. While that has helped power the S&P 500 (SNPINDEX: ^GSPC) and Nasdaq Composite (NASDAQINDEX: ^IXIC) to new all-time highs, it has also left both indexes very top-heavy with leading AI stocks.

For example, eight of the S&P 500's largest holdings are tech stocks that make up more than 35% of its holdings. Most of these are top semiconductor companies and hyperscalers (owners of large data centers). It's even more dramatic for the popular Nasdaq-100, whose 10 largest holdings are tech stocks and account for nearly half its portfolio.

Former Berkshire Hathaway CEO Warren Buffett. Image source: The Motley Fool.

At the same time, most of these stocks' fortunes are very tied to each other. Hyperscalers are spending a tremendous amount of money building out AI infrastructure, which, in turn, is driving growth at semiconductor stocks. If hyperscalers don't see a strong return on their AI investments and this spending stops, then these stocks could go down in unison, dragging down the major indexes with them.

One of investing stalwart Warren Buffett's favorite measures of whether stocks are overvalued is to take the total value of the U.S. stock market and divide it by gross domestic product (GDP). The former CEO of Berkshire Hathaway has said a range of between 75% and 90% is reasonable, while stocks start to look overvalued when it rises above 120%. This measurement, sometimes called the Buffett indicator, is currently over 235%, signaling a very expensive market.

Other popular metrics, like the cyclically adjusted price-to-earnings ratio, or CAPE ratio, also point to a frothy market. This metric smooths out earnings by adjusting for economic cycles, seasonality, and inflation over the past 10 years. Historically, this index's long-term average is around 17.4. It rose above 42 in August and is at its highest level since before the dot-com crash in 2000.

Now, whether the market has formed an AI bubble and whether the stock market will crash anytime soon is uncertain. However, with these indicators pointing to a frothy market, it is best to be prepared. Let's look at two things Buffett would recommend.

Have cash ready

One of Buffett's most famous pieces of advice is to "Be fearful when others are greedy, and greedy when others are fearful." Before he stepped down as CEO at Berkshire Hathaway at the start of this year, he had been following his own advice, not chasing stocks and instead letting his cash holdings climb.

One of the best things an investor can do before a potential market crash is to have some cash on the sidelines. This does not mean sell all your stocks and sit in cash. That would be a mistake, as no investor or market indicator can 100% predict when a market crash will occur.

However, if the market has a major pullback and you have no cash to invest, that also doesn't help. As such, if you want to follow Buffett, don't be fully invested and have some cash ready to pounce.

Buy companies with durable competitive advantages

Another important lesson investors can take from Buffett is which types of stocks to buy during a big market pullback. When the dot-com bubble burst, there were some great stocks to buy, but also many that never made it.

In an article in Fortune magazine, Buffett described the type of stocks investors should be looking for, saying, "The key to investing is not assessing how much an industry is going to affect society, or how much it will grow, but rather determining the competitive advantage of any given company and, above all, the durability of that advantage."

In other words, not all AI stocks will be long-term winners. Look for companies with wide moats that give them a long-term competitive advantage. Buffett seemed to leave one hint on an AI stock to buy if the market crashes, with his last big bet being on Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG), which is the most complete AI stock with both world-class chips and AI models to go along with a strong distribution and ad network moat.

Geoffrey Seiler has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet and Berkshire Hathaway. The Motley Fool has a disclosure policy.

Shares of SentinelOne (NYSE: S) sank after it reported its fiscal second-quarter results, despite the company surpassing revenue estimates and increasing its guidance. The stock remains up about 40% on the year despite the drop in its share price.

Let's take a closer look at its results and prospects to see if this is a chance to buy the cybersecurity stock on the dip.

Strong revenue growth and increased guidance

SentinelOne saw strong sales growth, with revenue climbing 21% in its fiscal Q2 to $292 million. That topped its earlier $289 million to $291 million forecast, and was ahead of the $290.3 million consensus. Adjusted earnings per share (EPS), meanwhile, doubled from $0.04 to $0.08 and were at the high end of its $0.06 to $0.08 guidance.

Annual recurring revenue (ARR), which is the annualized value of its customer subscription and consumption-based contracts, grew by 23% to $1.16 billion. SentinelOne added new net ARR of $56 million in the quarter, up 4%. Meanwhile, the number of customers with ARR of $100,000 or more rose by 17% to 1,702.

SentinelOne said it is seeing strong demand for artificial intelligence (AI) security, with both its Purple AI and Prompt Security solutions tripling their ARR year over year in the quarter. Purple AI acts as an advanced AI security analyst, helping users run complex threat hunts using only natural language queries. Prompt Security, meanwhile, helps customers govern and secure the AI tools and agents they are using. The company also saw its third straight quarter of cloud security ARR acceleration, while it said its data lake and AI SIEM (Security Information and Event Management) solutions are gaining traction.

Non-endpoint solutions now make up more than half its ARR, while its new Flex program was more than 10% of ARR. Flex is a flexible licensing model that helps customers add bolt-on solutions to core subscriptions more easily.

While it didn't give a specific number, the company said its dollar-based net retention rate improved year over year and expanded sequentially for customers spending $100,000 or more in ARR. Remaining performance obligations (RPOs), meanwhile, surged 45% to $1.7 billion. The company said the RPO growth is coming from both larger and longer contracts with customers.

Turning to guidance, the company projected fiscal third-quarter revenue of between $309 million and $311 million, which would equate to 20% growth. It expects adjusted EPS of between $0.08 to $0.09.

For the full year, SentinelOne upped its guidance. It now expects revenue to grow 20% to between $1.2 billion and $1.21 billion, with adjusted EPS between $0.30 and $0.32.

It said its opportunity has never been better, with accelerating momentum with both its core Singularity platform and its newer AI products. With $813 million in cash and no debt on its balance sheet, it continues to both invest in growth opportunities and buy back stock.

Image source: The Motley Fool.

Time to buy the dip

SentinelOne turned in a strong quarter and offered upbeat guidance, led by growth in its newer AI solutions. The cybersecurity industry as a whole is seeing good momentum after the so-called Mythos momentum, when Anthropic revealed that its new Mythos model was able to find previously undetected software vulnerabilities. Perhaps most promising was the company's RPO growth, although investors clearly wanted some more transparency, as the company doesn't give ARR guidance or current RPOs.

Meanwhile, the stock is much cheaper than its endpoint cybersecurity peers', trading at a price-to-sales (P/S) ratio of 6 times compared to 37 times for CrowdStrike and 22 times for Palo Alto Networks. That means that if revenue starts to accelerate, which looks quite possible given its RPO growth and the momentum it is seeing with its AI products, SentinelOne stock could really see a lot of upside from here.

Geoffrey Seiler has positions in SentinelOne. The Motley Fool has positions in and recommends CrowdStrike. The Motley Fool recommends Palo Alto Networks. The Motley Fool has a disclosure policy.

Shares of Marvell Technology (NASDAQ: MRVL) declined despite the company once again reporting strong data center and artificial intelligence (AI) revenue growth when it released its fiscal second-quarter earnings on Aug. 27. However, the stock is still up more than 150% year to date as of this writing.

Let's dive into the semiconductor company's latest results and prospects to see if this dip is a buying opportunity.

Robust data center revenue growth continues

Marvell has been a big beneficiary of the AI infrastructure build-out with both its connectivity and custom chip businesses. The company is a leader in optical DSP (digital signal processing) chips, which convert electrical data into optical signals for faster data transmission within data centers. This business is growing quickly as AI data centers move away from copper wiring to optical networks. It also has strong positions in broadband analog components and scale-out switching. It sees each of these businesses moving toward a $1 billion annual revenue run rate.

The company also has a strong custom chip business. Its IP (intellectual property) is used in Amazon's custom chips, and the cloud computing leader is currently its largest customer in this area. It's also involved with Microsoft's new Maia chip. However, the big buzz was about Marvell's recently announced partnership with Alphabet that includes inference accelerators, storage controllers, NICs, memory interface controllers, and near-memory compute. Marvell said the deal is broad-based and a game changer for the company, although it looks like it won't become a meaningful contributor until fiscal 2029 (calendar year 2028).

As for its results, its overall revenue jumped by 37% year over year to $2.74 billion, while its adjusted earnings per share (EPS) soared 40% from $0.67 a year ago to $0.94. Those results were just ahead of the midpoint of management's outlook for adjusted EPS of $0.93 on revenue of $2.7 billion.

Data center revenue jumped 46% year over year in the quarter to $2.17 billion. Communication and other end market revenue, meanwhile, rose 10% year over year to $567.8 million.

Looking ahead, Marvell management guided for fiscal 2027 Q3 revenue of $3.15 billion, plus or minus 5%, which represents year-over-year growth of about 52%. It is looking for adjusted EPS of $1.05 to $1.15. Third-quarter data center revenue is projected to surge by 75%.

It also upped its fiscal 2027 revenue growth outlook, taking it from $11.5 billion to $12 billion, representing 45% growth. Its data center business is now projected to grow 60%, up from a prior forecast of 50%. Data center growth is expected to be broad-based, with a significant acceleration in its custom chip business in the second half of fiscal 2027 and into fiscal 2028.

It is now projecting fiscal 2028 revenue to climb 50% to $18 billion, up from an earlier forecast of $16.5 billion. Its data center business is projected to grow by 60%, while its custom chip business is expected to more than double.

Image source: The Motley Fool.

Is it time to buy the dip?

Marvell has gone from a cheap stock, due to worries it was losing its lead partnership position with Amazon's custom chips, to an expensive stock riding a big optical interconnect wave. Even after this recent dip, the stock now trades at a forward price-to-earnings (P/E) ratio of under 34 times fiscal 2028 estimates (ending January).

The company's deal with Alphabet should kick in around the same time it loses any potential growth tied to future iterations of Amazon chips, which is a big win. Meanwhile, its optical opportunity is still in its relatively early stages and has the potential to be a huge growth driver. While I wouldn't jump on the stock right now, I do think it would become interesting on any further pullback.

Geoffrey Seiler has positions in Alphabet and Amazon. The Motley Fool has positions in and recommends Alphabet, Amazon, Marvell Technology, and Microsoft. The Motley Fool has a disclosure policy.

CrowdStrike (NASDAQ: CRWD) shares surged 20.5% on Aug. 27 after the company saw record new annual recurring revenue (ARR) and posted impressive metrics across the board. ARR is the annualized value of its high- gross-margin subscription contracts and does not include its professional services revenue.

With the share price of the cybersecurity company nearly doubling this year, the question is whether the stock still has room to run. Let's dig into the latest results and prospects to find out.

Impressive growth in the second fiscal quarter

With its strong fiscal Q2 results, CrowdStrike once again demonstrated why it is the premier endpoint security company. The rise of agentic AI is driving greater use of custom agents at endpoints and increasing risks, which in turn has made endpoint security a top spending priority for organizations. When Anthropic released its Mythos model, which could exploit previously unknown software vulnerabilities, demand for advanced AI-powered cybersecurity tools skyrocketed.

Image source: The Motley Fool

This increased demand is benefiting CrowdStrike, which said its endpoint business accelerated for the fourth straight quarter and is also fueling adoption of its next-gen security offerings. Among these offerings, AIDR (Artificial Intelligence Detection and Response) saw its ending ARR nearly triple versus Q1, while Identity and Next-Gen SIEM (Security Information and Event Management) ARR rose 39% to $2.1 billion. Cloud and Runtime Security ARR, meanwhile, jumped 29% to $905 million.

Its Falcon Flex licensing model, which lets customers access its complete cybersecurity product portfolio while deploying and paying for modules only when needed, also continues to be a differentiator. The subscription model saw its ARR double to $2.3 billion. It is seeing a more than 40% lift in ARR when existing customers convert to Flex, and a 53% increase when customers reFlex following their initial subscription. New customers are also starting off with Falcon Flex, accounting for 34% of its new Flex ARR.

Overall, CrowdStrike saw its revenue climb 26% year over year to $1.47 billion, with subscription revenue rising 27% to $1.4 billion. Total ARR increased 25% to $5.84 billion, with new ARR hitting a record of $332.8 million, up 51% year over year. CrowdStrike's adjusted earnings per share (EPS) soared 35% to $0.31. The results were ahead of the consensus for adjusted EPS of $0.29 on revenue of $1.44 billion.

Looking ahead, CrowdStrike raised its fiscal 2026 revenue guidance to between $5.99 billion and $6 billion from an earlier outlook of between $5.91 billion and $5.96 billion. It also increased its adjusted EPS guidance to a range of $1.25 to $1.26, up from a previous forecast of between $1.22 and $1.24 (adjusting for a 4-for-1 stock split).

For its fiscal third quarter, the company forecasted adjusted EPS of $0.31 on revenue of $1.52 billion to $1.53 billion. That was essentially in line with the consensus.

Is CrowdStrike stock still a buy after its huge run?

CrowdStrike delivered a very strong quarter, highlighted by exceptional new ARR growth. The company is seeing strong adoption of next-gen solutions across its cybersecurity portfolio, while its Falcon Flex flexible payment program continues to be a strong driver, helping it cross-sell new modules. With AI creating more cyber threats, CrowdStrike is well-positioned to continue seeing strong growth well into the future as the leader in endpoint security.

The biggest knock on the stock, though, is valuation. CrowdStrike shares trade at a forward price-to-sales (P/S) multiple of about 30 times analysts' estimates for the next fiscal year (ending January 2028). Given its overall mid-20% revenue and ARR growth, that's just a huge multiple.

As such, while I like the company and the opportunity ahead, I would not chase the stock here.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends CrowdStrike. The Motley Fool has a disclosure policy.

Salesforce (NYSE: CRM) shares surged 22.6% on Aug. 27 after the company reported solid fiscal Q2 results and issued upbeat guidance, driven by strong momentum from its agentic AI platform Agentforce, Data 360, and Slack.

However, the software-as-a-service (SaaS) stock is still down year to date and could have plenty of potential upside ahead, as it starts to dispel the narrative that AI will displace the software layer, and the stock remains cheap.

Let's take a closer look at its results and prospects and why the stock still looks like a buy.

Agentic AI continues to pick up steam

Agentforce continues to drive Salesforce's growth, with the AI agent platform's annual recurring revenue (ARR) surging more than 240% to $1.5 billion. The company's Slack platform has become one of its most important, with Slackbot its fastest-adopted AI product, seeing 150% sequential growth to 1 million active users after launching just five months ago.

To show that AI is not replacing the software layer, Salesforce said that nine of the 10 largest AI companies are using Salesforce and Slack, and their spending has increased by 435% year over year. It also introduced a new product called Claudeforce, a plug-in with pre-built sales skills built on Anthropic's Claude's reasoning and agentic tool use.

Data 360 (formerly Data Cloud), which helps customers unify their data into a single source, has also seen strong growth, as ARR tripled to $2.4 billion. Informatica added $1.1 billion in ARR, suggesting organic growth was over 60%. Combined Agentforce and Data 360 ARR climbed 210% year over year to $3.9 billion.

Overall, Salesforce's revenue jumped by 11% year over year to $11.35 billion, at the high end of its guidance range of $11.27 billion to $11.35 billion and above the $11.32 billion consensus compiled by LSEG. Subscription and support revenue increased by 12% to $10.82 billion. Much of the growth came from its Agentforce 360 and Slack platform, which saw revenue surge 43%.

Adjusted earnings per share (EPS) skyrocketed from 103% to $5.90. However, that included a $2.6 billion gain on strategic investments, largely from its investment in Anthropic. Excluding that gain, adjusted EPS would have been around $3.43, applying the company's tax rate to the gain, still well above the $3.27 consensus.

Looking ahead, the company once again increased its full-year guidance, as shown below:

Data source: Salesforce. Table by author.

For fiscal Q3, the company has forecast revenue to increase by 11% to 12% to between $11.42 billion and $11.5 billion. It is projecting adjusted EPS in a range of $3.42 to $3.44. Analysts were looking for adjusted EPS of $3.38 on revenue of $11.41 billion.

Image source: The Motley Fool.

The stock still looks like a buy

The growth that Salesforce is seeing comes from top AI companies, combined with the introduction of Claudeforce, which should help ease some of the fears about AI bypassing the software layer. The narrative never made much sense, and the sector's stocks have suffered greatly as a result, including Salesforce. However, the company has positioned itself very well for agentic AI, and growth is starting to show.

At the same time, the stock still looks attractively valued even after its big rebound. Based on next year's fiscal 2027 analyst estimates, it now trades at a forward price-to-sales multiple of 4 and a forward price-to-earnings (P/E) ratio of 16. For a stock with low-double-digit revenue growth and building momentum with Agentforce, it looks like a great GARP (growth at a reasonable price) stock to own.

Geoffrey Seiler has positions in Salesforce. The Motley Fool has positions in and recommends Salesforce. The Motley Fool recommends London Stock Exchange Group Plc. The Motley Fool has a disclosure policy.

Nvidia (NASDAQ: NVDA) left no doubt it remains the king of AI when the company reported its fiscal Q2 earnings after the bell on August 26. The chipmaker's revenue surged in the quarter, while its guidance came in well ahead of analyst expectations. It also did something it's never done before, giving a forecast more than a year out.

Let's dig into Nvidia's fiscal Q2 results and prospects to see why the semiconductor stock still remains a top AI stock to buy.

No slowing down

The law of large numbers is still not catching up with Nvidia, as the company's fiscal Q2 revenue growth accelerated to an eye-popping 106%, coming in at $96.2 billion. That's up from 85% growth in fiscal Q1, 73% growth in Q4 2025, 62% growth in Q3 2025, and 56% growth in Q2 2025.

Adjusted earnings per share (EPS), meanwhile, skyrocketed 120% to $2.22 from $1.01. The results easily topped analyst expectations for adjusted EPS of $2.09 on sales of $92.3 billion.

Data center segment revenue once again led the way, up 117% year over year to $89 billion. Nvidia's smaller edge computing segment, meanwhile, saw revenue rise 27% to $7.2 billion.

Within its data center segment, hyperscaler revenue soared 102% to $48.7 billion, helped by these companies' large AI infrastructure buildouts. AI cloud, industrial, and enterprise (ACIE) revenue, meanwhile, skyrocketed 138% to $40.3 billion. It said the growth was driven by neocloud expansions to meet rising enterprise, sovereign, and AI-start-up demand. It said it has nearly 20 customers that contribute $1 billion or more in annual recurring revenue.

Nvidia highlighted the strength of its ecosystem and new products as a big differentiator. Its networking business once again shone, with Spectrum-X Ethernet revenue growing 2.6 times year over year. Meanwhile, it expects its central processing unit (CPU) revenue to double in fiscal 2028, while Groq 3 LPX, its first rack-scale language processing unit (LPU) system built for inference, is in full production.

The company continues to throw off a huge amount of cash, with operating cash flow of $24.1 billion and free cash flow of $21.4 billion in the quarter. It ended the period with cash and marketable securities of $99.3 billion and $32.4 billion in debt. It also has $51.2 billion in non-marketable securities, which are its investments in non-public companies.

Looking ahead, Nvidia guided for fiscal Q3 revenue to come in around $108 billion, which would represent 89% growth. The growth will be led by the ACIE segment, with hyperscaler revenue picking up in fiscal Q4 and into fiscal 2028. What really shocked investors, though, was that the company projected revenue to grow by 70% in fiscal 2028, well above the 44% growth analysts have been projecting, while saying growth would be higher if not for supply constraints.

Image source: The Motley Fool.

The stock is still a buy

Nvidia continues to demonstrate why it's one of the best companies on the planet. It's much more than just a graphics processing unit (GPU) designer nowadays, becoming a complete AI infrastructure solutions provider with leading networking, CPU, and other chips that it can combine to create end-to-end AI server racks for specific AI tasks, including inference, agentic AI, and AI model training.

Notably, this quarter was the first time Nvidia has ever given a forecast a year in advance, speaking to the long-term visibility the company is seeing. CEO Jensen Huang also noted that demand is more broad-based and that the AI ecosystem is healthier than a year ago, as OpenAI is no longer the only force driving the AI infrastructure buildout, with multiple frontier model labs scaling up and a thriving open-model ecosystem, as well.

Despite its remarkable growth and increasing visibility, Nvidia's stock is still attractively valued, trading at a forward price-to-earnings (P/E) ratio of around 17 times based on the fiscal 2028 analyst consensus, with estimates likely to go up. Between its growth, valuation, and improving visibility, Nvidia is a must-own AI semiconductor stock.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Nvidia. The Motley Fool has a disclosure policy.

After a hot start to 2026, Broadcom (NASDAQ: AVGO) shares have lost significant steam. However, its Sept. 2 earnings report after the bell on Wednesday could prove to be a major catalyst. Let's look at five reasons to buy this AI stock ahead of earnings and for the long term.

The TPU king

Broadcom's biggest revenue driver is the custom AI chips it helped Alphabet develop, called Tensor Processing Units (TPUs). Alphabet is aggressively spending on AI infrastructure, raising its capital expenditure (capex) budget to between $195 billion and $205 billion, and indicating it will significantly increase it next year.

In addition, Alphabet has also begun to sell TPUs to customers outside of Google Cloud. Anthropic has placed $21 billion worth of TPU orders with Broadcom, and the three companies have signed an agreement for future TPU commitments. Broadcom delivers the chips and books the revenue, while Alphabet collects a licensing fee from Broadcom.

This all points to strong TPU growth for Broadcom in the coming years. And while other companies are starting to play a role with TPUs, Broadcom remains Alphabet's primary partner, especially for its flagship training chip.

Image source: The Motley Fool.

Jalapeño spices things up

TPUs are not the only custom chips that Broadcom has helped co-develop. OpenAI recently unveiled its new Jalapeño chip for inference, which Broadcom helped co-develop. In inference tests, the chip outperformed other AI chips, a result confirmed by SemiAnalysis. This includes Nvidia's Blackwell chips and TPUs, although it did not include Nvidia's newest Rubin chips, which are just starting to ship to customers now.

OpenAI will begin deploying Jalapeño chip in its AI data centers by the end of the year. This should become another big potential revenue boost for Broadcom in the future.

Access to the bottlenecks

If any hyperscalers want to design and make custom chips for inference or training at scale, they are going to need access to high-bandwidth memory (HBM) and manufacturing capacity. Broadcom has both. Broadcom has signed a $200 billion deal with Samsung Electronics to supply HBM, manufacturing, and advanced packaging through 2030. The company has also said that it has secured advanced node capacity with Taiwan Semiconductor Manufacturing.

The importance of this should not be overlooked. Nvidia has locked in much of SK Hynix's HBM capacity, with Amazon also a big customer, while Advanced Micro Devices has also secured capacity from Samsung. These companies also aren't relying on just one supplier, so that doesn't leave a lot of room for other players, and Micron Technology's HBM capacity is about a third of its Korean competitors. As such, having this access to HBM and manufacturing capacity is actually a huge moat in the custom chip business.

Optical upside

Given the importance of its custom chip business, it is easy to overlook that Broadcom is also well positioned to ride the shift to optical networking, which is faster and generates less heat than traditional copper wires, within AI data centers. Broadcom is a big player in the optical DSP (digital signal processor) market. DSPs are the primary chips used to take electrical signals from AI chip clusters and convert them into light, allowing the data to be moved through optical systems. It is also the leader in co-packaged optics, where it integrates optical receivers directly into its custom chips.

Optical connectivity is a fast-growing market, and this is another big potential area of growth for the company that plays directly into its custom chip business.

Attractive valuation

With the pullback in its shares over the past couple of months, Broadcom now trades at a forward price-to-earnings ratio (P/E) of just 18 based on analysts' estimates for fiscal 2027 (ending October 2027). That's a very attractive valuation given its growth prospects, and gives the stock plenty of room to run if it can deliver strong fiscal Q3 results and issue upbeat guidance in September.

Geoffrey Seiler has positions in Advanced Micro Devices, Alphabet, Amazon, and Broadcom. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Broadcom, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.

For investors seeking stocks with high yields and rising dividends, the energy midstream space is a great place to look. Two of the top stocks in the space are Energy Transfer (NYSE: ET) and Enterprise Products Partners (NYSE: EPD).

The two pipeline master limited partnerships (MLPs) have expansive midstream systems in the U.S. that handle different types of hydrocarbons, like natural gas, oil, and natural gas liquids (NGLs). They also both have attractive yields and have been increasing their distributions. Which stock is the better of the two to own, though, could largely come down to the type of investor you are.

Energy Transfer: The growth and dividend combo

Energy Transfer has one of the largest and most diverse midstream businesses in North America, with around 140,000 miles of energy infrastructure assets across the U.S. located in all major basins and connecting to major markets across the country. The company has never been afraid to chase growth, and it is currently in full growth mode, given the attractive opportunities it is seeing around natural gas.

The company has a very strong position in the Permian Basin, which is the United States' most productive and lowest-cost oil basin. It has also been home to some of the cheapest natural gas in the country due to pipeline takeaway constraints, although that is starting to ease with the start-up of new pipelines, including Energy Transfer's Hugh Brinson Pipeline. This pipeline, which recently came online, is one of the company's most important projects, as it can take natural gas from the Permian and supply it to markets throughout Texas and beyond, helping support the growing AI data center market. It also has another pipeline project that will take natural gas in the opposite direction to the Arizona and New Mexico markets.

In total, Energy Transfer expects to spend up to $5.9 billion on growth capital expenditures (capex) this year. These projects are all backed by long-term contracts, and the company expects to get mid-teen returns on its investments. Meanwhile, it is looking to raise its distribution at a 3% to 5% annual pace moving forward. With over 90% of its adjusted earnings before interest, taxes, depreciation, and amortization (EBITDA) coming from fee-based sources and a robust coverage ratio, the distribution is well supported, while the company's extensive system also allows it to frequently take advantage of market pricing dislocations.

Enterprise: The Steady Eddie

While Energy Transfer tends to take a more aggressive approach, Enterprise is conservative by nature. The company has very low leverage for the sector at 3 times, while also keeping a robust coverage ratio of 1.9 times last quarter. This approach has helped the company increase its distribution for 28 straight years through all types of energy markets and economic crises.

During periods of uncertainty, Enterprise has been quick to reduce its growth capex, taking it down to $1.8 billion and $1.6 billion in 2021 and 2022 after the COVID-19 pandemic. It plans to spend up to $4 billion this year, given the strong project opportunities it is seeing. While 2026 is a transitional year, Enterprise is projecting double-digit adjusted EBITDA and distributable cash flow (operating cash flow minus maintenance capex) growth in 2027.

The company's balance sheet is a major source of strength. In addition to its low leverage, it also has the highest credit rating in the midstream space and long-term debt at an attractive 4.7% weighted average cost. Meanwhile, Enterprise has been increasing its distribution at a 3% pace, which is something I'd expect to continue with the potential to boost that next year when its DCF growth will accelerate.

Image source: The Motley Fool.

The verdict

Whether Energy Transfer or Enterprise is the better stock to own really depends on the investor. For those looking for a steady, sleep-well-at-night investment, Enterprise is the clear choice given its track record. However, for investors looking for more potential upside, Energy Transfer is the better option. The stock is cheaper (forward enterprise value-to-EBITDA ratio of 8.3 vs. 10.7), has a higher yield (6.5% vs. 5.8%), and has more growth projects.

I own both and think you can't go wrong picking either. However, which one is the best to choice really depends what you're looking for. Of course, you can always own both.

Geoffrey Seiler has positions in Energy Transfer and Enterprise Products Partners. The Motley Fool recommends Enterprise Products Partners. The Motley Fool has a disclosure policy.

Dick's Sporting Goods (NYSE: DKS) shares collapsed following the company's fiscal second-quarter earnings report, as it warned that the athletic footwear and apparel space has become more promotional. In addition, it said there were fewer new footwear launches in Q2, and that those launches underperformed its own and industry expectations.

The commentary was similar to that of JD Sports Fashion, which operates stores under the Finish Line, Hibbett, and other banners. JD Sports also said it was facing a highly promotional environment and experiencing economic and footwear-product-cycle headwinds. Both Dick's and JD Sports are major Nike (NYSE: NKE) retailers, but industry headwinds could extend beyond Nike and affect other brands, including Lululemon Athletica (NASDAQ: LULU).

Let's look at why I'd be reluctant to buy these two apparel stocks even while they are down.

Nike

Nike is trading at its lowest level in more than a decade, but that is not enough reason to buy the stock. The company's former CEO, John Donahoe, who strained wholesale relationships and eschewed innovation in favor of relying on retro brands, left the company in a bad spot. Meanwhile, its "Win Now" initiative, begun in late 2024 under current CEO Elliott Hill, has yet to help meaningfully turn the company around.

While it's seen some positives, especially with running, overall, its struggles have continued. This is especially true in China, which was its strongest market before the pandemic.

However, the company has been losing share to local brands, and its sales and margins have collapsed. With the company trying to reestablish itself as a premium brand, the more promotional environment in North America will only add to its woes. It had already been closing stores in North America, which will likely pressure sales, so this just piles on top of that.

While Nike is an iconic brand, the company appears to have lost brand loyalty, and there is no guarantee that loyalty will return. Just look at Under Armour, which was once a hot brand that has now seen 13 straight quarters of revenue declines with no turnaround in sight. Meanwhile, trading at a forward P/E of 23, Nike's stock is still not cheap.

Image source: The Motley Fool.

Lululemon

While the warning from Dick's and JD Sports doesn't have the same direct impact on Lululemon as it does on Nike, there certainly appears to be a general shift away from sporting apparel and athleisure. That's bad news for a company like Lululemon that is already trying to turn around its business.

The company has already been struggling with product missteps and increased competition from newer high-end brands like Alo and Vuori. Meanwhile, it's been without a CEO for an extended period, and the hire of veteran Nike exec Heidi O'Neill, who will take over in September, disappointed investors, including activist investor Elliott Investment Management, which wanted turnaround specialist and longtime retail executive Jane Nielsen to lead the company.

China has been Lululemon's saving grace, with the company projecting 20% revenue growth this year. However, the company committed a major faux pas at a big yoga event held on the Great Wall when it accidentally gave a Chinese actor a Japanese taiko drum to play instead of a Chinese dagu drum. This huge PR misstep could certainly impact sales in what has been the company's best-performing market.

In July, Truist downgraded the stock on concerns that the brand was losing momentum, citing Google and TikTok search trends. With the overall sporting apparel and athleisure market showing troubling signs, the picture for Lululemon is likely only to get worse from here. Meanwhile, like Nike, the company is struggling to maintain its brand loyalty.

While Lululemon looks cheap at a forward P/E under 11, a low multiple on falling earnings isn't always a bargain, and the stock appears to be a value trap.

Geoffrey Seiler has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Nike, and Truist Financial. The Motley Fool recommends Lululemon Athletica Inc. and Under Armour. The Motley Fool has a disclosure policy.

Two top-performing artificial intelligence (AI) semiconductor stocks over the past year have been Advanced Micro Devices (NASDAQ: AMD) and Marvell Technology (NASDAQ: MRVL). AMD shares have climbed around 185% over the past year, while Marvell shares are up more than 227%.

Both stocks are riding hot trends, but one looks like the better chip stock to own over the next five years, in my view. Let's take a closer look at both.

Image source: The Motley Fool.

AMD: An inference and agentic AI winner

AMD is set to ride two of the biggest waves in AI right now in inference and agentic AI. While it was never able to challenge Nvidia in AI model training, the company is better positioned for inference, which isn't as technically complex and tends to be much more about quick access to memory than about raw computing power. On this end, the company's chiplet design allows its chips to be packaged with more memory, reducing latency.

The company has also made two major acquisitions in the memory space. Memory optimization company MEXT provides an AI solution that can direct less frequently accessed data from DRAM (dynamic random access memory) to flash memory and then retrieve it before it's needed, helping reduce costs without sacrificing performance.

Its acquisition of Taalas, meanwhile, gives it inference chips hardwired with specific AI models that can significantly reduce latency. It's also formed a partnership with Cerebras, in which its Helios system will be used in the pre-fill phase of inference, and Cerebras' high-end solution will be used for the decode phase.

AMD already has two large $100 billion inference deals in place with Meta Platforms and OpenAI, while Anthropic and Microsoft have also recently become customers. This should be a huge area of growth for the company over the next five years.

AMD also sees a $220 billion opportunity in server central processing units (CPUs). The company is a leader in this rapidly growing market, where demand has surged due to the need for more CPUs for agentic AI.

Marvell: An optics leader

Like AMD, Marvell is riding two powerful trends: optical networking and custom chips. Both businesses are growing quickly.

The company's strongest segment has been optical interconnects. AI data centers are beginning to move away from copper wiring to optical components, which generate less heat and offer higher speeds. Marvell, meanwhile, is the market leader in optical DSPs (digital signal processors), the primary chips that convert electrical signals from AI chip clusters into light, enabling data to be transmitted through optical systems. This business is booming, with the company expecting its interconnect revenue to climb 70% this fiscal year.

At the same time, Marvell is seeing strong growth in its custom chip business. It provides some intellectual property (IP) for Amazon's Trainium chips, which have been growing rapidly. The company also helped Microsoft develop its new Maia chip, and it will provide Alphabet with several components for its TPU architecture, including network interface controllers (NICs), AI inference accelerators, storage controllers, and near-memory compute. In addition, it has strong IP around SRAM (static random-access memory), which is used in Nvidia's LPUs (language processing units) and Cerebras' chips.

On the downside, Marvell may play a less prominent role in future iterations of Trainium, which could affect its growth. The stock is also expensive, trading at a forward P/E of 37 times fiscal 2028 (ending January 2028) analyst estimates.

The verdict

While both stocks look like potential long-term winners riding great trends, I think AMD should outperform over the next five years. It is slightly cheaper, trading at 30 times 2027 EPS, and doesn't face the potential Trainium headwind that Marvell does.

Geoffrey Seiler has positions in Advanced Micro Devices, Alphabet, Amazon, and Meta Platforms. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Marvell Technology, Meta Platforms, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

Billionaire investors were busy in Q2 buying shares of tech titans Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG), Apple (NASDAQ: AAPL), and Meta Platforms (NASDAQ: META). The three are among the largest companies in the world, with some of the most iconic tech businesses on the planet.

Let's dive into why these three growth stocks remain attractive investments that were recently bought by billionaire investors.

Image source: Getty Images.

Alphabet

Alphabet saw some big-name billionaire investors buy the stock in Q2, with the likes of Warren Buffett's Berkshire Hathaway and Renaissance Technologies' Jim Simons adding to their holdings during the quarter.

There are many reasons to like Alphabet. The company's core Google search business continues to have a large distribution moat through its ownership of the Chrome browser and Android smartphone operating system, as well as a revenue-sharing deal to be the default browser on Apple devices. Meanwhile, AI-powered tools, such as AI mode and AI Overviews, have increased queries and driven revenue growth.

In addition, the company's cloud computing business has been the fastest-growing of the big three cloud providers, with revenue surging 82% last quarter. The company also has a significant advantage in this area with its tensor processing units (TPUs), which give it a significant cost advantage over rivals that largely rely on graphics processing units (GPUs). The company has even started selling TPUs outside Google Cloud to select large customers, such as Anthropic, adding another high-margin revenue stream.

As the most complete AI company with top chips and models, Alphabet looks like a long-term AI winner.

Apple

Another stock that saw some billionaire buying in Q2 was Apple, with George Soros and Jeff Gendell of Tontine Partners both adding shares. While Apple's stock is not cheap, the company has arguably one of the best compounding business models on the planet.

The secret to Apple's business model is its closed ecosystem. Once a consumer buys an Apple device and starts taking photos, buying apps, and subscribing, they tend to get locked in. This, in turn, leads to greater use of high-margin Apple services, such as cloud storage, commissions on app purchases and subscriptions, and Apple Pay. The company also makes a tremendous profit from its revenue-sharing deal with Alphabet, under which it gets a 36% cut of search ad revenue generated by queries through its Safari browser.

Apple maintains a strong leadership position at the high end of the smartphone, PC, and device markets, all of which tend to have predictable replacement cycles. This creates a relatively repetitive revenue stream, while its services business continues to consistently compound at a high rate of return. Given this business model, Apple is a top stock to own long-term.

Meta Platforms

Meta was a popular buy among hedge fund billionaires in Q2, with Bill Ackman, Jim Simons, and Lee Anslie of Maverick Capital, among the buyers. It's easy to see why hedge funds were scooping up shares of the social media giant last quarter. The stock has been beaten down on concerns of its aggressive spending on AI infrastructure and is only trading at a forward P/E of 16 times 2027 analyst estimates.

Meta has been one of the best companies at harnessing the power of AI to drive growth at its core business. Its social media platform is a great flywheel business, where AI keeps users on the platform longer while also helping advertisers serve more relevant ads and achieve better conversions. This is leading to both more ad impressions and higher ad prices. Meanwhile, the company has been making great strides in its AI models, and CEO Mark Zuckerberg has big ambitions.

Geoffrey Seiler has positions in Alphabet and Meta Platforms. The Motley Fool has positions in and recommends Alphabet, Apple, Berkshire Hathaway, and Meta Platforms. The Motley Fool has a disclosure policy.

Memory stocks have been on a tremendous run over the past year, buoyed by surging prices, ballooning gross margins, and huge free cash flow. This has helped the stocks of Micron Technology (NASDAQ: MU), Sandisk (NASDAQ: SNDK), and SK Hynix (NASDAQ: SKHY) (before its U.S. IPO) go parabolic this year. Micron is up more than 700% over the past year, while Sandisk is up an almost inconceivable 3,400% during the same stretch. SK Hynix's Korean shares, meanwhile, have surged 600% during the past year.

The current memory supercycle can be tied directly to the AI infrastructure build-out, which has led to major supply-demand imbalances. The memory market is divided between DRAM (dynamic random access memory), used for very short-term storage, and NAND (flash), which retains data longer. Both have seen tremendous increases in AI-related demand.

The biggest driver in the market right now is high-bandwidth memory (HBM), which is packaged with graphics processing units (GPUs) and other AI chips to reduce latency and power consumption. It's one of the biggest bottlenecks in all of AI, and as such, the big three DRAM makers -- SK Hynix, Samsung, and Micron -- have been dedicating most of their resources to keeping up with demand. This is leading to a supply shortage across the entire DRAM market and skyrocketing prices.

Flash memory, meanwhile, is used in colossal enterprise solid-state drives (SSDs) that store training data. Flash has been supply-constrained as the big three memory makers have shifted their focus to HBM, after earlier cutting production and redirecting resources toward DRAM following the NAND market's crash after the pandemic. Stay-at-home mandates had led to a pull-through in demand for electronics, which are big users of flash memory, but once they were lifted, the market collapsed.

Image source: Getty Images.

History says the market will collapse, but this time does look different

History would tell us that memory stocks will eventually crash, as traditionally, boom cycles are met by increasing supply that eventually overtakes demand, causing memory prices to sink. However, this cycle does seem different.

First, AI infrastructure demand just continues to soar, with no signs of slowing down. Second, memory makers, especially those focused on HBM, are struggling to keep up, and a few factors will continue to constrain supply growth. This includes limited capacity growth for EUV (extreme ultraviolet lithography) machines needed in the manufacturing of both HBM and advanced logic chips, such as GPUs; HBM requiring upwards of three times the wafer capacity of regular DRAM; and the lead times required to construct new clean rooms. Finally, all of the memory makers, including pure-play flash maker Sandisk, have been able to secure long-term contracts for the first time in history.

With both DRAM and NAND memory cycles still looking to have a long runway ahead, and stock valuations not reflecting this, memory makers could continue to be top AI stocks to own.

Geoffrey Seiler 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.

One stock that has divided investors since its debut in June is Space Exploration Technologies (NASDAQ: SPCX). The Elon Musk-backed company had the largest IPO ever, and the stock has been volatile ever since, shooting up to over $225, then crashing to under $105 before settling in around its current price of roughly $135.

Let's dig into why this growth stock could skyrocket higher and also why it could eventually crash.

The bull case

SpaceX is a bet on the future and the vision of Elon Musk, who isn't afraid to make big, bold bets, and SpaceX has become his primary vehicle for that. Eventually, I would expect the company to acquire Tesla to put all his ambitions under one roof.

Perhaps the most important technological achievement the company has attained thus far is building reusable rockets. This greatly reduces launch costs and sets the company up for its future space aspirations. It is now working to take this technology to its massive Starship rockets, which would then open the door to putting data centers in space.

This is a potentially huge market for the company, and data centers in space would offer several advantages. The biggest is that they can be run by nearly continuous solar power. AI data centers are energy hogs, and putting them in space would help solve this major issue. It also would avoid local grid congestion and zoning issues, while bypassing congested fiber lines in favor of high-speed satellite optical laser links that could quickly transmit data across regions.

SpaceX's core terrestrial data center business currently has strong economics, with it saying it is getting a payback on its investments in less than a year. Meanwhile, its Starlink satellite internet business is a gem, giving it a strong recurring revenue business that is growing quickly. The company is also looking to potentially offer voice services to compete with traditional wireless carriers.

The biggest reason to own SpaceX, however, is the multitude of business opportunities it has and the industries it could disrupt.

Image source: The Motley Fool.

The bear case

While SpaceX has a lot of opportunities in front of it, it will also have to overcome a lot of technical hurdles. The company still needs to solve a multitude of engineering problems with Starship, including heat shield durability and catching the upper stage of the rocket with its launch tower. Being able to catch an expensive rocket and relaunch it shortly afterward is really the key to its business model.

If it can do that, it will then need to overcome the challenges of deploying data centers in space. One of the biggest hurdles is cooling systems in the vacuum of space, as the technology to do so has not yet been invented. Chips would also need to be developed that can withstand cosmic radiation, as high radiation levels can flip computing bits, corrupting them, and, over time, cumulative damage tends to degrade chips' performance. Meanwhile, a large-scale data center won't fit on even a mega-rocket, so an advanced robotic system would need to be developed to assemble it in space.

While Musk is undoubtedly a visionary, his track record with timely predictions is spotty at best. At the same time, what he is trying to accomplish will cost a considerable amount of money for a company that Morgan Stanley projects won't turn free-cash-flow-positive until 2035. As such, SpaceX is going to need to raise a lot of cash through debt and equity offerings to fund its capital expenditures.

Valuation is also another big issue. The stock trades at over 40 times price-to-sales (P/S), which is much higher than even Tesla traded at in its early days.

The verdict

Over the next few years, SpaceX is likely to trade more on investor sentiment and emotion than fundamentals, likely leading to a lot of volatility. That said, given its valuation, already massive market cap, and the capex-heavy nature of its pursuits, it's not a stock I'd buy at these levels.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Tesla. The Motley Fool has a disclosure policy.

Wolfspeed (NYSE: WOLF) had been one of the hottest stocks in the market this spring, surging on hopes that it could become the next AI winner.

The rise coincided with a bullish report from Substack publication Citrini Research, which had earlier come into prominence after publishing a thought piece about how artificial intelligence (AI) would negatively impact software-as-a-service (SaaS) stocks, helping sink that sector.

However, after its shares reached more than $80, Wolfspeed stock has come crashing back down to earth, retracing the big move it had made in May following Citrini pumping the stock. Its latest pullback coincided with another disheartening earnings report on Aug. 19.

Image source: The Motley Fool

Negative gross margins and weak sales persist

Wolfspeed emerged from bankruptcy last fall, wiping out some expensive debt and finding itself on better footing. However, many of the issues that pushed it into bankruptcy in the first place remain. The chief among them is negative gross margins.

Wolfspeed positioned itself as the leader in silicon carbide (SiC) powered chips. The company constructed expensive manufacturing plants to build out a vertically integrated supply chain. SiC has superior heat-conducting properties compared to typical silicon chips, and thus initially was projected to play a major role in the electric vehicle (EV) market.

However, the company ran into severe execution bottlenecks and market headwinds. The move to larger 200mm wafers proved to be more technically challenging than imagined, while EV demand started to slow. Meanwhile, Tesla decided to greatly reduce its use of SiC moving forward.

That left Wolfspeed with severely underutilized, brand-spanking-new plants, which is one of the worst things a semiconductor company can experience. It is also one of the reasons why most traditional silicon-based chipmakers use a fabless model and rely on third-party foundries like Taiwan Semiconductor Manufacturing.

Wolfspeed's operational issues persisted in its fiscal fourth quarter, with the company seeing negative gross margins of 25% due to continued plant underutilization. Adjusted gross margins, meanwhile, came in at negative 19.9%, a 70-basis point sequential improvement.

Revenue growth continues to be an issue, with revenue falling 24% year over year from $197 million to $149.6 million. It was also a slight sequential decline from $150.2 million in fiscal Q3 and right in the middle of its $140 million to $160 million outlook.

While auto revenue remained soft, the company did see its AI data center revenue more than double year over year and rise 20% sequentially. It highlighted a few new design wins with power supply companies, while saying SiC content was increasing in data centers due to the transition to 800-volt architectures.

Given its negative gross margins and $600 million in net debt, cash flow remains an issue. The company has negative operating cash flow of $180.3 million over the past nine months after exiting bankruptcy, and negative free cash flow of $253.7 million. It will look to retire some high-interest debt to help with its cash flow moving forward.

Looking ahead, the company once again guided for quarterly revenue to be in the $140 million to $160 million range and for gross margins to remain negative. It said it would likely need revenue to reach $800 million in revenue for its adjusted gross margins to break even.

Is the stock a buy on the dip?

Even after coming out of bankruptcy, Wolfspeed still finds itself in a precarious position. Negative gross margins and free cash flow are never great signs for a business, and the company will need to see a meaningful acceleration in sales just to get to breakeven gross margins, which isn't exactly a huge accomplishment.

That said, the transition to 800-volt architectures in AI data centers is a legitimate structural shift, and the opportunity for SiC is both real and accelerating. So, there is the potential for a big turnaround in the stock. This makes Wolfspeed an interesting, but highly speculative, stock to buy at this point.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Taiwan Semiconductor Manufacturing and Tesla. The Motley Fool recommends Wolfspeed. The Motley Fool has a disclosure policy.

Shares of Walmart (NASDAQ: WMT) saw their worst one-day drop since 2022 after the retailer said it would use its approximately $2.9 billion in tariff refunds to lower prices for consumers, which would negatively impact its third-quarter results.

The company indicated that consumers appear to be stretched due to high gasoline prices, but that overall they remain resilient and keep spending, helped by wage growth.

Following its dip, the stock is now down about 7% on the year, as of Aug. 20. Let's dive into the retail stock's quarterly results to see if the pullback is a buying opportunity or if investors should stay away.

Image source: Getty Images

Solid results but soft outlook

Walmart has always been known as the low-cost leader, so its decision to push down prices in response to the tariff refund isn't completely surprising. However, lowering costs will impact its sales growth and operating margins in the near term.

For Q3, Walmart forecasted that its sales would rise between 3% and 3.75%, which was well below the nearly 6% consensus. Its adjusted EPS guidance of $0.62 to $0.64, meanwhile, fell short of the $0.68 that analysts were expecting.

Despite the disappointing Q3 guidance, the company did raise its full-year outlook. It now anticipates adjusted EPS of between $2.80 and $2.87 on sales growth of 4% to 5%. That's up from a prior projection for EPS of $2.75 to $2.85 on revenue growth of 3.5% to 4.5%. Ultimately, the retailer expects its decision to lower prices to help accelerate sales and lead to market share gains.

Turning to Walmart's fiscal Q2 results, its revenue rose nearly 6% to $187.94 billion, topping the $186.77 billion consensus, as compiled by LSEG. Walmart U.S. store sales rose 3.5% to $125.2 billion, while same-store sales increased by 2.6%. The number of transactions rose by 1.5%, while the average ticket climbed 1.1%. E-commerce sales, meanwhile, surged 24%. Excluding health and wellness, which is being impacted by industry regulations, same-store sales would have been 3.4%

Internationally, Walmart sales climbed nearly 13% to $35.2 billion, and were up nearly 8% in constant currencies. China led the way with nearly 21% constant currency growth. International e-commerce sales jumped 19%.

Sam's Club U.S., its warehouse store concept, saw sales (excluding fuel) increase by 4.5%. Same-store sales, excluding fuel, climbed 4.4%. Transactions rose 7%, while the average ticket fell 2.5%. E-commerce sales jumped 26%, while membership revenue rose 6% year over year.

Adjusted EPS climbed 19% to $0.81. That topped the $0.74 consensus.

Is the stock a buy on the dip?

Walmart is one of the world's great retailers, and the company has proven to be a solid grower in all types of economic conditions, which helps set it apart in the space. In recent years, the company has done a great job of bringing in more affluent customers, catering not just to their desire for low prices, but also goods like higher-quality meats and produce, and same-day delivery convenience. As such, even as its core lower income customers have shown some struggles, Walmart continued to prosper.

While the decision to invest its tariff refund in price cuts will impact short-term results, it once again should create strong loyalty among its customers. That's a pretty good long trade-off.

That said, the stock was pricey going into the earnings report, and it still isn't cheap after the sell-off, trading at a forward price-to-earnings (P/E) ratio of nearly 36 times current fiscal year estimates and nearly 32 times next year's fiscal estimates. I think Walmart remains a solid long-term holding and that it will eventually bounce back, but I much prefer the cheaper and faster-growing Amazon stock in the retail space.

Geoffrey Seiler has positions in Amazon. The Motley Fool has positions in and recommends Amazon and Walmart. The Motley Fool recommends London Stock Exchange Group Plc. The Motley Fool has a disclosure policy.

Legendary investor Warren Buffett loves investing in great compounding businesses with wide moats at attractive valuations. While he has largely avoided tech stocks in the past, that has changed in the last few years, with him leading the charge with investments in Apple and Alphabet.

Let's look at three tech stocks that would be great fits for Warren Buffett's portfolio, even if he is unlikely to buy the stocks anytime soon.

Image source: The Motley Fool

Meta Platforms

Few companies know how to monetize free users better than Meta Platforms (NASDAQ: META), and that is only getting better through the use of AI. The company has the perfect flywheel business model for AI, where its investments in improved models are keeping users on its sites longer and helping advertisers better convert them to customers. This is turn is leading to both more ad impressions and higher ad prices due to the increased effectiveness of the ads.

With over 3.5 billion daily active users from around the globe, Meta enjoys massive global network effects and scale efficiencies. The company produces massive operating cash flow that it can now invest in high return investments in AI, keeping the flywheel going. Meanwhile, the stock is on sale, trading at a forward P/E below 16 times 2027 analyst estimates.

The one thing likely keeping Buffett from owning the stock, though, is CEO Mark Zuckerberg's penchant for wasting money with Reality Labs. However, this is a stock that's still a buy in my book.

Microsoft

Microsoft (NASDAQ: MSFT) has a huge enterprise lock-in with Office 365, which gives it both solid pricing power and the vehicle to provide AI tools and agents to a huge customer base. With trusted security and guardrails in place, Microsoft has been seeing strong adoption of its Microsoft 365 AI assistant Copilot, which has been driving strong growth in its enterprise software segment.

Meanwhile, Buffett has already shown a fondness for the economics of cloud computing, given his investment in Alphabet. Microsoft Azure shares similar characteristics with strong, steady revenue growth. The company also has one of the largest backlogs in the space given its relationship with OpenAI.

Even after a recent jump in price, the stock is still attractively valued, trading at 24.5 times this fiscal year's analyst earnings consensus. It also owns a 27% stake in OpenAI, which could become the next $1 trillion IPO.

The one thing that likely keeps Buffett away from the stock is his soured relationship with Microsoft founder Bill Gates. However, the stock still looks like an attractive buy.

Taiwan Semiconductor Manufacturing

The arms dealer in the AI infrastructure boom, Taiwan Semiconductor Manufacturing (NYSE: TSM) has an unassailable competitive moat built around its engineering expertise and scale. As the only foundry that has proven to produce advanced logic chips at high yields at scale, the company has created a virtual monopoly in the space.

This has given the company huge customer lock-in and pricing power. It's become a vital cog in the semiconductor industry and an important partner to chipmakers, which basically have to coordinate with TSMC on their technological roadmaps and capacity expansion plans. While semiconductor manufacturing is a capital-intensive business, this ensures that the company gets strong returns on its investments.

Buffett has called TSMC "one of the best-managed companies and most important companies in the world," while it has also garnered praise from Nvidia CEO Jensen Huang, who called it "one of the greatest companies in the history of humanity, and anybody who wants to buy TSMC stock is a very smart person." That's some high praise, yet the stock remains cheap, with a forward P/E of just 19 times 2027 analyst estimates.

While Buffett was briefly invested in the stock in the past, geopolitical tensions have kept him from making it a long-term investment. However, that seems like a long-shot black swan event, and those fears should not keep investors from buying this great stock.

Geoffrey Seiler has positions in Alphabet and Meta Platforms. The Motley Fool has positions in and recommends Alphabet, Apple, Meta Platforms, Microsoft, and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.

When it comes to artificial intelligence (AI) semiconductor stocks, Nvidia (NASDAQ: NVDA) is the gold standard, and there is still a lot to like about the stock. It remains the dominant chip company for AI model training, and that is unlikely to change. Most early foundational AI code was written using its CUDA software platform and optimized for its graphics processing units (GPUs), creating a wide moat in this arena.

Meanwhile, Nvidia hasn't rested on its laurels. The company is a huge player in data center networking, which has formed the base for it to become a complete AI infrastructure player offering end-to-end rack-scale solutions designed for specific AI tasks. The company also created its own custom Arm-based central processing units (CPUs), which will serve it well as the rise of agentic AI drives servers focused on AI agents to contain more CPUs.

Nvidia has also positioned itself well for the inference market through its "acquisition" of Groq and its language processing units (LPUs). LPUs contain a small amount of SRAM (static random-access memory) and are strung together to help reduce latency and speed up inference speeds. For systems designed for inference, Nvidia will combine its GPUs, which will handle the compute-heavy pre-fill phase of processing a prompt, with LPUs, which will provide low latency for the decode phase of generating a response.

Nvidia continues to produce incredible growth, with its fiscal first-quarter revenue soaring 85% to $81.6 billion while its adjusted EPS surged 140% to $1.87. Despite that, the stock is attractively valued, trading at a forward P/E of about 24.5 times fiscal 2027 (ending January 2027) analyst estimates. Given that, I think the stock still looks like a buy. However, given its size and the law of large numbers, I predict another AI semiconductor stock, Advanced Micro Devices (NASDAQ: AMD) can outperform it over the next three years.

Image source: The Motley Fool

AMD: Ready to take on Nvidia in inference

While Nvidia controls the market for large language model (LLM) training, AMD is set to be a serious competitor when it comes to inference. Importantly, this is the faster-growing market and also expected to become the much larger one. According to projections from Bloomberg Intelligence, the inference market is projected to grow at a 32% compound annual growth rate through 2032 and become nearly double the size of the AI model training market, hitting annual spending of $1.3 billion.

Inference is less technically demanding than AI model training, and AMD has done a great job improving its ROCm software stack over the past few years. Meanwhile, inference is generally more memory-bound than compute-constrained, and the company's chiplet design, which can be packaged with more memory and is designed to act as part of an entire system, reduces latency.

AMD also made two acquisitions in the space to better position it for the inference market. First, it acquired memory optimization company MEXT, which offloads less-frequently used data from DRAM to underused flash memory and then uses predictive AI to transfer it back before it's needed. It then bought inference chip start-up Taalas, which hardwires AI models directly onto its chips to bolster inference performance. It uses specific AI models, so it isn't as flexible, but it's a cheap and fast inference alternative.

Along the same lines, AMD also partnered with Cerebras to offer a disaggregated system for inference. AMD's Helios system will deal with the pre-fill phase of inference, while Cerebras' ultrafast but expensive solution will handle the decode phase. It is a partnership that should benefit both companies.

At the same time, AMD also has a huge opportunity with agentic AI. The company is a leader in server CPUs, and has been steadily taking share in this market away from Intel. With the ratio of GPUs to CPUs moving from 8:1 for training to 1:1 for agentic AI, AMD sees this becoming a $220 billion market in the next few years.

As a much smaller company than Nvidia, AMD really has the opportunity to see explosive growth and for its stock to outperform over the next three years as it makes serious inroads in the inference and agentic AI markets. With big deals already in place and its revenue growth about to take off, I think the stock can outperform Nvidia over the next three years.

Geoffrey Seiler has positions in Advanced Micro Devices. The Motley Fool has positions in and recommends Advanced Micro Devices, Arm Holdings, Intel, and Nvidia. The Motley Fool has a disclosure policy.

When it comes to artificial intelligence (AI) and software-as-a-service (SaaS) stocks, investors have generally taken a cautious approach. Between AI disintermediation, vibe coding, and the potential shift to more consumption-based business models, investors have identified several risks that could impact the industry.

However, not all SaaS companies are alike, and it is more likely that software companies will play a critical role in the future of AI. One of the big reasons for this is that they are model-agnostic. If AI models shift in leadership or move more to commoditization, organizations don't want to be tied to a single model. At the same time, software is often deeply embedded within organizations and has important compliance and security guardrails in place, making this a strong option for AI to be deployed.

Against that backdrop, let's look at three SaaS stocks with big upside from AI.

Image source: Getty Images

Palantir Technologies

When it comes to benefiting from AI, no SaaS company has been better at it than Palantir Technologies (NASDAQ: PLTR). It has taken its roots as a data-gathering and analytics company for the U.S. government and created one of the most important AI platforms on the planet.

In simple terms, its AI platform essentially acts as an AI operating system that significantly decreases potentially harmful AI hallucinations and makes AI more useful to organizations in the real world. The strength of the platform can be seen in the company's surging revenue growth and remarkable net dollar retention, which came in at 157% in Q2.

While the stock is not cheap, Palantir has the potential to grow to become one of the largest companies in the world. AIP can be used to help solve problems across various departments and industries, making its growth runway just enormous.

Microsoft

Among megacap tech stocks, Microsoft (NASDAQ: MSFT) has been one of the most underappreciated. The company remains the 800-pound gorilla of enterprise software, and its Microsoft 365 suite of productivity tools remains deeply entrenched within enterprises.

Big organizations tend to be cautious about new technologies like AI, so they've been turning to a trusted leader like Microsoft to help them incorporate AI into their workflows. This is leading to strong adoption of Microsoft's AI assistant, Copilot, and is helping drive strong growth in its enterprise software segment. This should remain a strong growth driver for the company.

At the same time, Microsoft's cloud computing unit, Azure, has been growing rapidly and has one of the largest backlogs in the space. The company is also just starting to place greater emphasis on developing its own custom AI chips and foundational AI models, which should help improve margins over time. Throw in a 27% stake in OpenAI, and this is a SaaS stock with a lot of potential AI upside.

GitLab

Turning to a much smaller SaaS company, GitLab (NASDAQ: GTLB) has all the ingredients to see its stock surge higher. The stock is cheap, trading at an attractive price-to-sales multiple of just 5.5 times based on fiscal 2028 (ending January 2028) analyst estimates, amid investor concerns about the impact of AI on its DevSecOps (development, security, and operations) platform.

However, GitLab looks much more likely to be an AI winner than a loser. AI is driving more software development, not less, and all of it still needs to be done in a secure environment. Meanwhile, commentary from Microsoft, which owns code repository GitHub, and JFrog, which provides software supply chain solutions in the DevSecOps sector, point to strength in the sector. Microsoft also noted growth being driven by its move to a hybrid consumption model, which is also something that GitLab has moved toward with its Duo Agent Platform.

Given its valuation and AI-driven growth, GitLab stock has big potential upside from here.

Geoffrey Seiler has positions in GitLab. The Motley Fool has positions in and recommends Microsoft and Palantir Technologies. The Motley Fool recommends GitLab and JFrog. The Motley Fool has a disclosure policy.

Quantum computing has the potential to be the next big game-changing technology after artificial intelligence (AI), and the race is on for companies to develop and commercialize a fault-tolerant system. It's a potential breakthrough technology that could help create the next round of tech giants.

While many companies are pursuing quantum computing, a few stand out above the rest. These companies are ahead of the pack largely because of their technology leadership. One of the first big obstacles with quantum computing that needs to be overcome is accuracy.

Image source: Getty Images

Because quantum computers use qubits rather than classical fixed bits, they are very sensitive to external factors, such as vibrations and temperature changes, that can throw them off and cause errors. In the world of computing, being correct 99% of the time over billions of calculations is extremely error-prone and not usable at scale. Quantum hardware will likely never be 100% perfect on its own, but it doesn't have to be. To work reliably, quantum computers need fault tolerance, a built-in safety net that constantly catches and corrects hardware errors as they occur, yielding virtually 100% accurate results.

Here are the three quantum stocks that are ahead of the pack in achieving a fault-tolerant quantum system that could change everything.

IonQ

When it comes to quantum computing accuracy, IonQ (NYSE: IONQ) is the current leader. Its trapped ion approach starts with actual atoms, which are identical in nature and thus more stable. However, instead of entirely relying on complex lasers, the company has embedded microwave antennas directly into its chips to control the qubits electronically and improve stability. The result is that IonQ has achieved 99.99% two-qubit gate fidelity (accuracy), the top mark on any public company.

However, IonQ has not stopped there, and it is looking to control the entire quantum ecosystem. It has made acquisitions in several related areas, including quantum networking, transmission, and sensing. Its most important deal to date, though, could be its acquisition of quantum foundry SkyWater to become completely vertically integrated. By having its own foundry, IonQ should be able to advance prototypes more quickly and eventually use it to scale its systems.

Given this, IonQ is the quantum stock to beat.

Quantinuum

Similar to IonQ, Quantinuum (NASDAQ: QNT) also takes a trapped-ion approach, but it has eschewed the use of microwave antennas, claiming that microwave gates are too slow and that attempts to speed them up increase power consumption excessively. The company has achieved 99.92% 2-qubit gate fidelity with its systems, which trails IonQ, but it is looking for its new Sol system to reach 99.999% logical fidelity in 2027. That would be a major achievement.

The Honeywell-backed company is also highly regarded for its leadership in quantum software. It has developed an open-source quantum computing language called Guppy that manages advanced error correction directly in the code. Meanwhile, its TKET quantum software development kit acts as a universal translator, and its Nexus platform brings everything together in a cloud-based operating system.

If there is a company set to beat IonQ to the punch, it is Quantinuum.

Infleqtion

Infleqtion (NYSE: INFQ) doesn't yet have the accuracy metrics of IonQ or Quantinuum, but its neutral-atom approach is intriguing. Similar to the trapped-ion technique, it uses individual atoms suspended in a vacuum that are controlled by lasers. However, it keeps the atoms uncharged, so that they don't repel one another. This, in turn, lets it attain higher qubit density and could make the technology more scalable than the trapped-ion approach.

The goal is to create a system with comparable accuracy to trapped-ion systems with faster speeds. It reached 99.73% 2-gate fidelity in 2024 and is expected to reach 99.9% this year.

If Infleqtion can catch up in fidelity metrics, its speed advantages could put it in the driver's seat.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Honeywell Technologies and IonQ. The Motley Fool has a disclosure policy.

Shares of Marvell Technology (NASDAQ: MRVL) shot higher following the news this week that Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG) would extend a partnership with Marvell related to its tensor processing units (TPUs). Meanwhile, Broadcom (NASDAQ: AVGO), which is Alphabet's main design partner for TPUs, saw its shares sink on the announcement.

As part of the deal, Marvell will provide Alphabet with a variety of components within its TPU architecture. These include things like AI inference accelerators, storage controllers, network interface controllers (NICs), and near-memory compute. In return, Marvell has given Alphabet warrants to buy 58.97 million shares of its stock at a price of $206.65 per share. The warrants vest in tranches based on every $500 million the cloud computing and search giant spends with it.

Despite the deal, Broadcom still looks like the better buy among these two semiconductor stocks.

Broadcom

Broadcom helped Alphabet develop its TPUs more than a decade ago, and it has been the company's main chip design partner ever since. Earlier this year, it signed a contract with the company to remain its primary design partner through 2031. The two companies also later agreed to a deal to supply Anthropic with TPUs, as well.

The Marvell deal will not impact this, but Alphabet has clearly been adding other partners to its TPU ecosystem. For upcoming TPU iterations, the company has already decided on two versions of its chips, with one for high-performance training and another version for inference. Broadcom is the main partner for the former, while MediaTek has taken the lead for the inference chips. Given the components involved, Alphabet's deal with Marvell should impact MediaTek more than Broadcom.

It's also been rumored that Alphabet is working with Advanced Micro Devices, capitalizing on its central processing unit (CPU) expertise for another future TPU version for agentic AI workloads. If true, Alphabet is looking to really broaden its TPU ecosystem. That ultimately should be good for Broadcom, as it is still a major supplier of networking components, like high-bandwidth Ethernet and optical interconnects.

At the same time, Broadcom still has a huge TPU opportunity in front of it, and it is helping other major AI data center players including Meta Platforms and OpenAI to develop their own custom chips. Broadcom management expects its custom chip business to contribute over $100 billion in revenue in fiscal 2027, and anticipates that business continuing to grow strongly in the year to come.

Trading at a forward P/E of around 18.5 times fiscal 2027 estimates, Broadcom stock looks like a bargain given its expected growth.

Image source: The Motley Fool

Marvell

Marvell has been one of the hottest large-cap tech stocks in the market over the past year, with its shares more than tripling. However, its performance has largely been due to excitement around its connectivity business, as optical interconnects are starting to replace copper wires within AI data centers. It projects its interconnect revenue will surge 70% this year, with overall revenue climbing 40%.

The company's custom chip business has also been strong, largely due to its role in Amazon's Trainium chips. However, there has been speculation that it will lose its lead role with Amazon to Taiwanese semiconductor company AIchip in future iterations. This deal with Alphabet, along with one with Microsoft for its new Maia chip, could help replace any lost Amazon revenue or compensate for reduced growth elsewhere.

Marvell is a solid stock, and the company's optical interconnect business should continue to see strong growth. However, the stock reached a pretty frothy valuation, trading at a forward P/E ratio of 57.5 and 37 times next year's expected earnings.

With Broadcom trading at a much cheaper valuation and still anticipating strong growth that should not be materially impacted by the Marvell-Alphabet agreement, the stock looks like the better buy in my view.

Geoffrey Seiler has positions in Advanced Micro Devices, Alphabet, Amazon, Broadcom, and Meta Platforms. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Broadcom, Marvell Technology, Meta Platforms, and Microsoft. The Motley Fool has a disclosure policy.

Billionaire investor Paul Tudor Jones is best known for correctly calling the 1987 market crash, making an estimated $100 million in the process. The legendary investor has publicly struck a cautious tone about the market, and during the second quarter, he was making moves around his top AI holdings, including Nvidia (NASDAQ: NVDA), Micron Technology (NASDAQ: MU), Microsoft (NASDAQ: MSFT), and Taiwan Semiconductor Manufacturing (NYSE: TSM).

Jones tends to use a combination of holding common shares along with call and put options. Let's look at how he was repositioning his top AI holdings in the second quarter.

Nvidia

Nvidia is Jones' third-largest stock position and the one AI stock it appears he made the most bullish moves around, significantly increasing his common stock holdings, while only slightly adding to his put position. However, he did significantly reduce his call position.

The combination of Nvidia's valuation (the stock trades at a forward price-to-earnings (P/E) of just 17 based on analyst estimates for its fiscal 2028, ending January 2028) and growth (revenue grew 85% in the first quarter) makes it an attractive stock at current levels. As the AI chip leader, Nvidia still has a huge opportunity in front of it, and its acquisition of Groq positions it well for the booming inference market. However, its future clearly revolves around overall AI infrastructure spending, which Jones is clearly hedging against.

Micron Technology

Jones' most bearish moves around his top AI holdings were with memory maker Micron, as he slashed his call positions and lowered his common stock exposure while upping his put position.

Micron has been riding the memory supercycle, which has led to surging revenue and ballooning gross margins. While the cycle looks like it could last several more years and the stock is cheap with a forward P/E under 7, this is historically an industry that has seen boom times lead to eventual collapses.

While industry supply-demand dynamics suggest this time will be different, as memory makers struggle to keep up with growing demand as they deal with their own constraints, Jones' heavily hedged position does make sense.

Taiwan Semiconductor Manufacturing

Taiwan Semiconductor is Jones' largest stock position. He increased his number of common shares by 340% in the quarter, while more than doubling his put position. He also added slightly to his call position.

TSMC is arguably one of the companies best positioned for the AI infrastructure boom. The company has a virtual monopoly on advanced logic chip manufacturing, making it a vital partner for chip designers. It's benefiting from increased chip demand, while also seeing strong pricing power. Best of all, it doesn't matter which AI chip company prevails, as most rely on it for their chip manufacturing.

That said, TSMC is also the AI company with the most to lose. If AI infrastructure spending drops and it's left with a lot of underutilized fabs, that would spell disaster for its margins. As such, Jones' heavily hedged position makes sense.

Image source: Getty Images.

Microsoft

Of his top four AI holdings, Microsoft appears to be the one Jones is overall most bullish about, with his common stock and call holdings larger than his put position. During Q2, he significantly raised his common stock and put positions by pretty similar amounts, while drastically cutting his call position.

Microsoft stock had gotten pretty beaten up in Q2, as investors feared that AI could disintermediate its core software business. They were also cautious about its cloud computing unit's heavy ties to OpenAI. However, with its fiscal fourth-quarter earnings report in July, Microsoft continued to prove that AI is a growth driver for its core enterprise software business, while its Azure cloud business continues to boom.

Given how entrenched Microsoft is in the enterprise space, the company looks like it will be an AI winner, not a loser. Meanwhile, it has a huge cloud backlog that will help drive growth in this segment for many years to come.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Micron Technology, Microsoft, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.

AI chip stocks have been volatile this year as investors weigh the enormous amount of spending going toward AI infrastructure and how sustainable it is. However, with cloud computing companies seeing strong returns on their AI chip and networking investments with quick payback periods and locked-in contracts, it appears that this robust spending will continue.

Let's look at three AI semiconductor stocks that are off 15% or more from their highs to buy right now.

1. Broadcom

Down 23% from its highs set this spring, Broadcom (NASDAQ: AVGO) looks like one of the most attractive chip stocks to buy on the dip. The company is a leader in data center networking and custom AI ASICs (application-specific integrated circuits) and has a significant growth opportunity ahead.

Broadcom helped Alphabet develop its powerful Tensor Processing Units (TPUs), which are set to be a major growth driver for the company. Alphabet is spending aggressively on growth capital expenditures (capex) this year and has indicated it plans to spend significantly more next year. This should feed directly into Broadcom's TPU and networking business. Meanwhile, Alphabet has even let Anthropic place TPUs directly with Broadcom, adding another growth driver.

Given the success of TPUs, other hyperscalers have also turned to Broadcom to help them create their own custom AI chips. Broadcom has projected it will see more than $100 billion in ASIC revenue in fiscal 2027, while Citigroup estimates that will rise to $180 billion in fiscal 2028.

With the stock trading at below 20 times fiscal 2007 analyst earnings estimates, the stock is a buy on this pullback.

2. AMD

Advanced Micro Devices (NASDAQ: AMD) has been a hot stock in 2026, but the recent chip pullback has left it about 18% off its highs. That makes it a great time to jump into a stock that is riding two of the biggest trends in AI: inference and agentic AI.

After losing to Nvidia in AI model training, AMD is making sure it will grab a piece of the larger and faster-growing inference segment. It formed two large partnerships with OpenAI and Meta Platforms centered around inference, which helped give AMD a strong foothold in this market. At the same time, through its chiplet design, which can package more memory, and deals to acquire memory optimization company MEXT and inference chip company Taalas, AMD is aggressively looking to position itself as a leader in this market. It also smartly teamed up with Cerebras for a disaggregated system where its graphics powered unit (GPU)-powered Helios solution will handle the pre-fill phase more cheaply, with Cerebras' more expensive technology reducing latency.

On top of that, as a leader in server central processing units (CPUs), AMD is also set to ride the wave in agentic AI. AI agents are creating a huge need for advanced CPUs, and AMD sees this becoming a $220 billion market in the next few years. With the company riding both these trends, it's time to buy the dip.

Image source: Getty Images.

3. SK Hynix

Down around 20% from its high following its initial public offering (IPO) this year, SK Hynix (NASDAQ: SKHY) is a top memory stock to grab on the pullback. The Korean company is one of the big three DRAM makers and the market share leader in high bandwidth memory (HBM).

HBM is currently the driving force in the memory market, as GPUs and other AI chips require this specialized form of DRAM (dynamic random-access memory) to reduce latency and optimize performance. However, a combination of factors, including HBM requiring upwards of three times the wafer capacity and ordinary DRAM, is keeping capacity tight, while demand continues to grow.

With long-term deals in place and as the main supplier of HBM to Nvidia, SK Hynix looks like the best-positioned memory maker over the long term. It sees the market being imbalanced until at least 2030, although there is a good chance this supercycle lasts much longer. With a forward price-to-earnings (P/E) around 6 times, the stock looks like a buy on this dip.

Citigroup is an advertising partner of Motley Fool Money. Geoffrey Seiler has positions in Advanced Micro Devices, Alphabet, and Broadcom. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Broadcom, and Nvidia. The Motley Fool has a disclosure policy.

Bill Ackman is one of the world's most renowned investors. The billionaire hedge fund manager runs a very concentrated portfolio of fewer than 15 stocks, so when he makes moves, investors take notice. While he added new positions in credit card networks Visa and Mastercard, financial intelligence company S&P Global, and video streamer Netflix, those weren't his only notable moves.

During the quarter, he made some big moves with stocks of hyperscalers (owners of large data centers). Ackman finished selling off his remaining small Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG) stake, while slashing his position in Amazon (NASDAQ: AMZN) by 25%. He also added to his positions in Microsoft (NASDAQ: MSFT) and Meta Platforms (NASDAQ: META).

Let's take a closer look at these hyperscaler stocks to see if investors should follow suit.

Microsoft

Microsoft is Ackman's third-largest holding, and he upped his position by nearly 10% in the second quarter. It's easy to see why the legendary investor likes the stock. Even after a recent jump, the stock has done very little over the past year despite continued strong operational performance.

The worry is that AI will disrupt its core software business, while some investors have questioned the economics of AI-focused cloud computing in general and Microsoft's heavy ties to OpenAI. However, Microsoft's software is deeply embedded in enterprises, and the increasing adoption of its AI copilots shows that the company is becoming an AI winner.

Meanwhile, its Azure cloud computing unit continues to put up outstanding growth and has a huge backlog. This is a stock to own for the long term.

Meta Platforms

Meta is a top growth stock trading at a very attractive valuation, with a forward price-to-earnings (P/E) of just 15 times 2027 analyst estimates. While investors have frowned at its AI infrastructure spending, few companies have been as good at employing AI to drive growth in their core business as Meta.

Ackman likes the company for its great flywheel business, where continued AI improvements to its algorithms lead to users staying on its apps longer and advertisers being better able to connect with them. This is leading to both increasing ad impressions and increasing prices, which is driving strong growth for the company. Meanwhile, Meta has started to make some impressive strides with the release of its recent AI models.

This is a stock to buy while it remains cheap.

Amazon

While Ackman cut his stake in Amazon by a quarter, it remains his fifth-largest position. However, this is not a stock I'd personally be looking to trim. The company is currently firing on all cylinders, and the stock remains relatively cheap both from a historical standpoint and compared to some of its brick-and-mortar retail peers, like Costco Wholesale and Walmart.

The company's e-commerce business just continues to deliver solid revenue growth. More impressively, it's seeing tremendous operating leverage from its internal investments in AI and robotics, as well as its high-margin sponsored ad business. At the same time, its cloud computing business is seeing accelerating revenue growth, backed by partnerships with Anthropic and OpenAI. Amazon is spending aggressively to increase capacity, and its custom chip business helps give it a cost advantage.

The company is getting a strong locked-in return on these investments, and management thinks Amazon Web Services can become a $1 trillion revenue business in the future. This is a stock I'd continue to buy at current levels.

Image source: Getty Images.

Alphabet

After he cut most of his position in the first quarter, it was not surprising to see Ackman completely exit Alphabet in Q2. However, this is a stock I really like.

The company is the most complete AI player, and it has several big advantages. This includes its custom Tensor Processing Units (TPUs), which are the most powerful custom AI chips out there, giving it a big cost advantage for inference and training its AI models. It also has a huge distribution edge through its ownership of Chrome, Android, and its search revenue-sharing deal with Apple, while its powerful global ad network is helping it monetize AI better with consumers than most AI model companies.

Alphabet is well-positioned to be an AI leader over the long haul, and it's a stock I'd want to continue to own.

Geoffrey Seiler has positions in Alphabet, Amazon, and Meta Platforms. The Motley Fool has positions in and recommends Alphabet, Amazon, Apple, Costco Wholesale, Mastercard, Meta Platforms, Microsoft, Netflix, S&P Global, Visa, and Walmart. The Motley Fool has a disclosure policy.

Billionaire investor David Tepper of Appaloosa Management has made a name for himself as one of the world's top investment minds. The hedge fund manager is worth an estimated $23.7 billion, and he was recently busy selling high-flying memory stocks while adding to positions in several Magnificent Seven names.

This includes Amazon (NASDAQ: AMZN), Meta Platforms (NASDAQ: META), Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOGL), and Nvidia (NASDAQ: NVDA). Let's dive into why Tepper likely likes these stocks.

Amazon

Tepper's largest position is in Amazon, representing over 15% of his portfolio, and he was adding more shares in Q2. It's easy to see why Tepper would like the stock. Amazon is both the market leader in cloud computing and e-commerce, and it's been showing strong growth in both areas.

The company's cloud computing unit, AWS, has been seeing accelerating growth, with revenue climbing 37% year over year in Q2, its fastest growth in four and a half years. With a huge backlog, partnerships with Anthropic and OpenAI, and the company spending aggressively on high-return AI infrastructure projects, the strong growth should continue.

Amazon also has an important custom chip business that is growing quickly, while it also uses the chips to help reduce its own internal inference costs.

On the e-commerce side, the company is seeing great operating leverage come from its internal investments in robotics and AI. It's also developed a large high-margin digital ad business that continues to grow quickly. Between its cloud computing and e-commerce opportunities, Amazon is a top stock to own.

Meta Platforms

Tepper was also aggressively scooping up shares of Meta Platforms in the quarter. While some investors have been concerned about the company's AI infrastructure spending, it looks like one of the best values among megacap growth stocks. It has a forward P/E of 16 times 2027 analyst estimates, while it just grew its revenue by 28% in the second quarter.

The company has one of the best flywheel business models out there for AI, with AI advancements helping both keep users on its sites longer and advertisers better reach and convert them into customers. Meta has also made strong progress with its newest AI models and is looking to become an important overall AI player. In addition, the company has a long runway of growth from it just starting to introduce ads to its Threads and WhatsApp platforms.

Alphabet

Alphabet is Tepper's fourth-largest holding, and he was adding to the position in Q2. Alphabet is the most complete AI company, having both its own top-tier chips and AI models.

Its biggest advantage is its tensor processing units (TPUs), which it developed more than a decade ago and which it has built its entire hardware and software stack around. This gives it a cost edge for internal use and a nice margin boost from customers that opt for them with Google Cloud. It is also starting to let some large select customers like Anthropic deploy the chips in their own data centers.

Meanwhile, Alphabet's core search business continues to grow nicely, with AI-powered features, like AI Mode and AI Overviews, helping drive query growth. It also owns one of the largest streaming services in the world, YouTube, and has a big emerging opportunity with its robotaxi business Waymo.

A great set of emerging and leading businesses, Alphabet is a solid core holding.

Image source: Getty Images

Nvidia

During Q2, Tepper also added to his Nvidia position, which is another top-10 holding. Nvidia has been the king of AI infrastructure, with its graphics processing units (GPUs) the main chips used to train AI models. Meanwhile, the company also has a nice opportunity in the inference market, where the combination of its GPUs and the language processing units (LPUs) it attained through its "acquisition" of Groq, gives it a unique end-to-end solution.

With the stock cheap, trading at a forward P/E of 17.5 times fiscal 2028 (ending January 2028) analyst estimates, and continuing to grow rapidly, Nvidia looks like a nice buy at current levels.

Geoffrey Seiler has positions in Alphabet, Amazon, and Meta Platforms. The Motley Fool has positions in and recommends Alphabet, Amazon, Meta Platforms, and Nvidia. The Motley Fool has a disclosure policy.

Chase Coleman III of Tiger Global Management is one of the world's premier tech investors. In fact, his fund has spawned many other top investors who now run their own funds. During Q2, Coleman was busy, including reducing his stake in Nvidia (NASDAQ: NVDA), while adding a new position in Cerebras Systems (NASDAQ: CBRS) and increasing his holdings in Intel (NASDAQ: INTC).

Let's take a closer look at these semiconductor stocks and see if investors should be buyers or sellers.

Image source: Getty Images

Nvidia

While Coleman reduced his stake in Nvidia by about 7%, it remains his third-largest holding. There are many reasons to like the stock here, in my view. First, the company is the dominant chipmaker for AI model training with its graphics processing units (GPUs), and that is unlikely to change anytime soon. Its CUDA software platform, where most foundational AI code has been written, just remains a powerful moat.

Meanwhile, the company has become much more than just a GPU maker. Backed by a strong networking portfolio and other chips, Nvidia is now a complete AI infrastructure player that can provide end-to-end rack-scale solutions for various AI tasks. Its "acquisition" of Groq and its language processing units (LPUs) give it a strong position in the fast-growing inference market, while its custom central processing units (CPUs) set it up well for agentic AI.

Nvidia continues to see explosive growth and has big opportunities still ahead of it, yet the stock trades at a forward P/E of just 17 times fiscal 2028 (ending January 2028) analyst estimates. That makes this a great stock to buy at these levels in my view.

Cerebras

Coleman's biggest buy in the quarter was Cerebras, as he made the chip company his 12th-largest holding. Cerebras is one of the most intriguing chip stocks in the market.

Similar to Nvidia's LPUs, Cerebras embeds SRAM (static random-access memory) directly onto its chips, which greatly increases inference speeds. However, SRAM is bulky, and instead of using a small amount like LPUs, it has created massive wafer-sized chips that need special cooling and power management. The result is speeds up to six times faster than LPUs. However, this comes at a cost, and it is generally viewed as a premium solution.

Cerebras has started to become a player at the high-end of the market, and its system is helping power the new ultrafast mode for OpenAI's GPT-5.6 Sol model. The company has a large commitment in place with the AI model maker that should provide strong growth in the years to come. It also has a deal with Amazon Web Services.

Meanwhile, the company's partnership with Advanced Micro Devices also looks promising to help push the company's solution more into the mainstream with a more cost-effective option. Through the partnership, AMD will provide its Helios system to handle the pre-fill phase of inference, while Cerebras' solution will power the decode phase. The result will be a cheaper, low-latency inference solution that benefits both chipmakers and lets them better compete with Nvidia in this huge and fast-growing market.

Cerebras is a more speculative stock, but the ingredients are in place for it to be a big winner.

Intel

Last quarter, Coleman significantly increased his stake in Intel, making it his 15th-largest holding. The company has been a late AI winner, with the rise of agentic AI leading to a surge in demand for data center CPUs. The company remains the CPU market share leader, so it has seen a nice jump in revenue and an improvement in gross margins.

While Coleman finds the stock attractive, it is not a favorite of mine. The company has been losing share in the server CPU market to AMD, while Arm Holdings has entered the market to add more competition. Meanwhile, it still appears to be trailing in technology and is benefiting more from the high tide raising all ships. On top of that, higher component costs will likely negatively impact sales in its core PC segment, while its foundry business continues to see heavy losses.

The stock is up big over the past year and no longer cheap, and as such, I'd stay on the sidelines.

Geoffrey Seiler has positions in Advanced Micro Devices and Amazon. The Motley Fool has positions in and recommends Advanced Micro Devices, Amazon, Arm Holdings, Intel, and Nvidia. The Motley Fool has a disclosure policy.

Stanley Druckenmiller, the billionaire fund manager of Duquesne Capital, is one of the world's most preeminent investors. So when he makes portfolio moves, people take notice. Among the many moves he made in the second quarter, he closed his positions in Micron (NASDAQ: MU) and Intel (NASDAQ: INTC), while opening new stakes in Advanced Micro Devices (NASDAQ: AMD) and Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG).

Let's take a closer look at these AI stocks to see if investors should follow suit.

Micron and Intel

Given the performance of Micron and Intel's stocks this year, Druckenmiller made some hefty profits in these positions. However, one stock certainly looks like a better option to keep holding than the other.

The stock I'd be more willing to hold is Micron. The company has been riding the memory supercycle, as supply-and-demand imbalances have caused memory prices to skyrocket. This, in turn, has led to huge surges in revenue and gross margins for Micron.

While the memory market has historically been highly cyclical, the massive AI data center infrastructure build-out has changed the market dynamics. In order for graphics processing units (GPUs) and other AI accelerators to deliver optimized performance, they need to be packaged with large quantities of high bandwidth memory (HBM), but a combination of factors is set to keep the market supply constrained for years.

Meanwhile, with the big three memory makers all focused on increasing their HBM production capacity, the entire DRAM (dynamic random access memory) market is seeing a huge price increases. Yet Micron's stock is trading at a cheap forward price-to-earnings (P/E) ratio of just over 6.5. With the memory supercycle set to potentially last several more years, the stock looks like a buy.

With Intel, on the other hand, I would take profits and not look back. The company is riding a wave of rising demand for data center central processing units (CPUs) as hyperscalers and neoclouds prepare for an extended surge in the use of AI agents, but this appears to be more the company stumbling into good fortune rather than turning its fortunes around. Meanwhile, its foundry business continues to be a money-losing drag. With the stock no longer cheap, I'd remain on the sidelines.

Image source: Getty Images.

AMD: Riding two big trends

While Druckenmiller dumped Intel, he didn't abandon the server CPU theme; he added a stake in AMD. AMD is the leader in the CPU market, having consistently been taking share from Intel. Meanwhile, its high-core CPUs are designed specifically to handle agentic AI workloads. The company sees this becoming a $220 billion market in the coming years and believes it can win more than 50% of that market.

AMD also has a big opportunity in the AI inference market, which is expected to become much larger than the market for AI training. The company's chiplet design enables it to package its processors with more memory, which is particularly beneficial for inference workloads. It also formed a partnership with wafer-scale engine specialist Cerebras to offer a disaggregated system designed specifically for inference. AMD's recent acquisitions of chipmaker Taalas and memory optimization company MEXT also set it up well for this market.

With huge opportunities stemming from agentic AI and inference, AMD looks poised for explosive growth in the coming years, making the stock a solid buy.

Alphabet: The complete AI player

Alphabet may be the most complete AI play, so it's easy to see why Druckenmiller made it one of his top 10 stock holdings in Q2. The tech giant has been seeing huge growth in its cloud computing segment, while its custom AI chips, called Tensor Processing Units (TPUs), give it a big cost advantage for inference workloads.

These chips also let it train its AI models more cheaply -- models that it then incorporates throughout its products, including Google Search, to drive growth. Alphabet also has a big distribution edge through its ownership of the Chrome browser, the Android operating system, and a search revenue-sharing deal with Apple that makes Google the default search engine on its devices. Meanwhile, its global digital ad network helps it better monetize consumer AI than most large language model (LLM) makers.

As the company with the most complete AI stack, Alphabet looks like it will be a long-term AI winner.

Geoffrey Seiler has positions in Advanced Micro Devices and Alphabet. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Apple, Intel, and Micron Technology. The Motley Fool has a disclosure policy.

Billionaire tech investor Philippe Laffont of Coatue Management was busy adding stocks to his portfolio in the second quarter. According to the 13F form filed on Aug. 14, the largest new addition was Space Exploration Technologies (NASDAQ: SPCX), although that likely stemmed from pre-IPO investments. Meanwhile, he aggressively bought a few semiconductor stocks in the quarter, including Micron (NASDAQ: MU), Cerebras (NASDAQ: CBRS), and Intel (NASDAQ: INTC).

Investing alongside billionaire investors can sometimes be a smart strategy, but should retail investors follow Laffont into those stocks?

Image source: The Motley Fool.

SpaceX

Like most large fund managers, Laffont likely acquired his SpaceX shares in the private market at valuations much lower than the stock's IPO price or its current trading level. However, based on interviews, it does not appear he plans to exit the position anytime soon, despite likely carrying a large gain.

Laffont has said it's easy to dismiss SpaceX as overvalued, and while he doesn't know what the company should be worth today, he thinks space is going to be a huge opportunity over the next 10 to 15 years. However, color me skeptical. I don't really want to invest in a company when there's no good way to value it except based on its 10- to 15-year potential. While the pace of technological innovation is picking up, there are still many technological hurdles the company will need to overcome. Those issues are enough to keep me on the sidelines.

Micron

After adding a small amount of Micron shares to his portfolio in Q1, Laffont dramatically increased the number of shares in his position by almost 19-fold in Q2. While Laffont didn't catch the memory trade early in the up cycle, it appears he is betting that it is here to stay.

This makes sense in my view. Between the finite number of EUV (extreme ultraviolet) lithography machines that ASML can produce, the fact that high bandwidth memory (HBM) uses upward of three times the wafer capacity of ordinary DRAM (dynamic random access memory), the constraints on advanced chip packaging capacity, and the time it takes to build new clean rooms for chip production, it looks like memory makers will not be able to boost their output by enough to catch up with surging demand for quite some time.

Micron stock is trading at a cheap forward P/E ratio of just above 6.5, and the memory supercycle looks as if it has legs. I like the stock here.

Intel

Intel's stock has been on a huge run over the past year, so Laffont is getting to this party late as well. While Intel largely missed out on benefiting from the initial artificial intelligence boom, its revenue has finally started to climb as demand for server central processing units (CPUs) surges due to the expected growth of agentic AI.

While data centers built in recent years for AI model training were outfitted with around 8 GPUs to 1 CPU, in servers that will handle agentic AI, the required GPU-to-CPU ratio is 1 to 1.

Because of that, there has been a steep increase in demand for data center CPUs. However, Intel has been losing market share in this space to Advanced Micro Devices, and Arm Holdings is also making a big push. At the same time, Intel's foundry business continues to bleed cash. Now that its valuation has gone from cheap to expensive over the past year, I wouldn't be chasing Intel stock.

Cerebras

Cerebras is a third semiconductor stock that Laffont was aggressively buying in Q2. While it's a more speculative investment, this is a stock I can get behind. The company's wafer-scale engines are unique, but with large quantities of SRAM (static random-access memory) directly embedded in them, they can offer superior inference performance, albeit at a premium price.

The company has already shown it may become an important player in the high end of the inference market. It has a large deal with OpenAI, and it is powering the Ultrafast mode of its new GPT-5.6 Sol model. At the same time, a recent partnership with AMD, whose Helios system can provide a more cost-effective solution for the pre-fill phase of inference, could also help Cerebras break into the mainstream segment of the inference market. That makes the company an intriguing potential long-term winner in a huge and fast-growing market.

Geoffrey Seiler has positions in Advanced Micro Devices. The Motley Fool has positions in and recommends ASML, Advanced Micro Devices, Arm Holdings, Intel, and Micron Technology. The Motley Fool has a disclosure policy.

After Cerebras Systems (NASDAQ: CBRS) reported its second-quarter results after the close on Aug. 12, its shares sank 16% over the following two trading sessions, although the stock bounced back after the investment advisor Wedbush Securities praised the company for powering OpenAI's ultrafast mode for its GPT-5.6 Sol model.

The chipmaker's stock is now down by around 43% from the high it touched shortly after its initial public offering earlier this year, and the stock looks like a buy as inference demand soars.

Cerebras makes systems built around its wafer-scale chips -- processors the size of dinner plates that are made from a whole silicon wafer. They can contain a large amount of static random-access memory (SRAM), which is an advantage, but they also require special cooling and power management, which makes them a premium option.

However, having many standard chips' worth of hardware on a single extra-large chip also makes them super-fast and ideal for the decode phase of inference. The company is teaming up with Advanced Micro Devices, whose Helios solution will cut down costs and offer strong value. It also has big deals in place with OpenAI and a partnership with Amazon through Amazon Web Services that is expected to go live early next year.

Soaring revenue and improving gross margins

Cerebras saw 74% revenue growth in the second quarter, with sales climbing to $180.1 million. Its core sales figure, which strips out revenue distortions caused by customer warrants and pass-through accounting, more than doubled to $209.9 million.

More and more customers are renting out its systems, which led to its cloud revenue surging 281% year over year to $126 million, and its core cloud revenue soaring 287% to $127.7 million. Hardware revenue sank 23% year over year to $54.1 million, while core hardware revenue rose 17% to $82.1 million.

Gross margins have been a point of contention for the company. Its core gross margins came in at 40.6%, up 940 basis points versus a year ago. Core gross margin for its cloud operation was 41.8%, a 1,600 basis point year-over-year improvement, while core hardware gross margin was 38.8%, 510 basis points higher than a year earlier. However, core gross margin fell sequentially from 46.5% in the first quarter due to higher costs from renting back systems to meet urgent demand.

Management projected third-quarter revenue of between $214 million and $216 million, with core gross margins between 38% and 40%. Cerebras also upped its full-year guidance, taking its core revenue forecast to a range of $880 million to $890 million, up from a prior outlook of $855 million to $865 million. It now sees its core gross margins coming in between 41% and 43%, up from an earlier projection between 38% and 41%.

For 2027, management is looking for core revenue to surge more than threefold and is expecting strong growth in 2028 and beyond. It is projecting core gross margins to improve in 2027 and move toward its 60%-plus long-term target.

Image source: The Motley Fool.

Time to buy the stock

The AI inference market is heating up, and the size of that segment is expected to eventually become much larger than AI training for Cerebras. Bloomberg Intelligence is projecting it will grow at a 32% compound annual rate through 2032 to reach $1.3 trillion, nearly double the size of the AI training market. While Nvidia dominated the AI processor market when it came to hardware for training, the inference market looks like it will have multiple winners, including Cerebras.

Although the company's systems are a more expensive option, their superior performance should help it gain its fair share of this rapidly growing segment. The OpenAI announcement is the perfect example of the opportunity in front of Cerebras at the high end of the market.

Meanwhile, I really like its partnership with AMD. A combined solution, where AMD chips can more cheaply handle the pre-fill inference phase -- the initial stage of large language model inference -- should be very compelling and help give customers the best of both worlds. Cerebras' deals with OpenAI and Amazon also provide a strong core customer base, and over time, its customer base should expand.

Given the growth of the inference market and the huge opportunity ahead of it, Cerebras looks like a solid buy, albeit a speculative one, on its recent price dip.

Geoffrey Seiler has positions in Advanced Micro Devices and Amazon. The Motley Fool has positions in and recommends Advanced Micro Devices, Amazon, and Nvidia. The Motley Fool has a disclosure policy.

With second-quarter earnings now in the books for quantum computing stocks, it appears that two are starting to pull away from the pack: IonQ (NYSE: IONQ) and Quantinuum (NASDAQ: QNT). This perhaps should not be surprising, as these are the two companies using a trapped-ion approach, which has thus far proven to be the most accurate. IonQ has reached 99.99% two-qubit gate fidelity, while Quantinuum has achieved 99.92%, putting them both far ahead of the pack in this metric.

This edge in accuracy is also starting to show up in their earnings results.

Image source: Getty Images.

IonQ: Surging revenue

IonQ's Q2 revenue soared 287% to $80.1 million, which was well ahead of the $66.4 million average estimate. Importantly, 60% of its revenue came from commercial, non-government customers, showing its solutions are moving beyond lab experiments. Multi-product sales, meanwhile, jumped 40% and accounted for about a quarter of its revenue.

Its order backlog rose to $485 million, up from $122 million a year ago, and it raised its full-year revenue forecast to $280 million to $290 million, excluding its recently closed SkyWater acquisition. The acquisition of the foundry is expected to accelerate its quantum roadmap, as it looks to eventually develop 10,000-qubit chips by 2027. The company also highlighted its move from lasers to its proprietary Electronic Qubit Control (EQC) technology, which uses microwave antennas built directly on its chips. This will help it scale as it lowers costs and reduces energy consumption.

Quantinuum: Oracle partnership is a game changer

Quantinuum's Q2 revenue surged 279% to $8 million, driven by growth in its cloud business, but the highlight of its earnings report was its strategic partnership with Oracle. Its Helios system will be integrated with Oracle's cloud infrastructure to give customers a quantum-artificial intelligence framework.

The company's order backlog, meanwhile, climbed to $74 million and is projected to reach at least $120 million by year-end. Meanwhile, Quantinuum is looking for its new Sol in 2027 to reach 99.999% logical fidelity. It projects 2026 revenue between $28 million and $32 million, and more than doubling in 2027.

The rest of the pack

The other quantum names largely reported mixed results. Infleqtion's (NYSE: INFQ) revenue jumped 116% to $12.6 million, and it raised its full-year outlook to $43 million. However, its backlog was up only slightly sequentially at about $21 million. The company has a more mature quantum sensing business, while its neutral-atom technology, which is similar to trapped-ion but with the charge removed, holds promise. It is looking to demonstrate 30 logical qubits on its system by the end of 2026.

D-Wave Quantum (NASDAQ: QBTS), known for its annealing systems, reported revenue of $3.1 million for the quarter, little changed. It expects Q3 revenue to be similar to Q2, before seeing a big jump in Q4. Importantly, the company is getting into full-fledged gate-based quantum computers, and it said its superconducting dual-rail qubit architecture hit 99.9% two-qubit fidelity. While it showed great speed, in the world of computing, that's a very wide accuracy gap compared to IonQ and Quantinuum.

Rigetti Computing (NASDAQ: RGTI) reported revenue growth of 185% year over year to $5.1 million. However, like D-Wave, its system also trails significantly in accuracy. Its Cepheus-1-108Q currently operates at a median two-qubit gate fidelity of about 99.1%, while its nine-qubit system has achieved 99.8%. While its systems are fast, it really needs to make big strides in accuracy to be considered a serious contender in the quantum race.

Accuracy is winning

Based on their backlogs and revenue, IonQ and Quantinuum are showing that in the race to quantum supremacy, accuracy is more important than speed. IonQ is the leader, while Quantinuum looks poised to make a big leap. That makes these two the stocks to own in the segment right now.

Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends IonQ and Oracle. The Motley Fool has a disclosure policy.

Investor Michael Burry, who made a name for himself by successfully profiting from the housing market collapse, has been one of the biggest bears when it comes to investing in artificial intelligence (AI) stocks. He holds short positions in several large tech stocks, including chip giant Nvidia, memory maker Micron, AI operating platform Palantir, and Elon Musk's Tesla. In a bet against the entire AI infrastructure trade, he is also short the iShares Semiconductor ETF.

However, Burry took his negative views to a new level recently, calling for a potential stock market crash reminiscent of the one from 1987. Known as Black Monday, the Dow Jones Industrial Average, which was widely considered the benchmark index at the time, dropped a record 22.6% on Oct. 19, 1987. A combination of a long bull market, rising interest rates, international tension in the Persian Gulf (sounds familiar), the start of computerized trading, and hedging strategies that led to more selling are often cited for the crash.

In a post on Substack, Burry wrote: "I continue to believe it is possible we are near a major top, and [possibly] a 1987-type fall, but the S&P 500 making new highs likely will bring new money into the market. ... Remember, the market going up on falling volatility forces vol-targeting funds to leverage up, and brings leverage from other momentum strategies into play."

Despite Burry's success with the housing collapse, a 1987-type crash seems unlikely to happen again. The introduction of circuit breakers, or mandatory trading halts during severe market drops, has helped stem panic selling and prevented a similar crash since. Meanwhile, hyperscalers aren't slowing down. Rapid payback periods on AI chips and networking hardware, coupled with locked-in customer agreements, continue to justify their aggressive infrastructure spending. At the same time, most AI stocks don't look expensive.

Against this backdrop, I think the best strategy for the average investor is to stick to dollar-cost average into index exchange-traded funds (ETFs). Consistently buying index ETFs through bull and bear markets has historically proved to be a great wealth-building strategy. One of the big reasons for this is that market-cap-weighted indexes basically deploy a survival of the fittest strategy that lets their winners run and losers fade, which has worked wonders over the long term. Even as AI infrastructure stocks were pulling back earlier this year, the S&P 500 index was still holding strong as new leaders, like Apple, emerged.

Image source: Getty Images

Let's look at two top index ETFs I'd continue to dollar-cost average into right now.

The Vanguard S&P 500 ETF

Few actively managed funds have been able to consistently beat the S&P 500 Index over the long term. As such, one of the smartest decisions someone can make is to invest alongside the index by dollar-cost averaging into a low-cost ETF such as the Vanguard S&P 500 ETF (NYSEMKT: VOO).

The ETF consists of 500 of the largest U.S. stocks, giving investors an instant portfolio of top companies. The ETF has been a strong performer over the years, generating an average annual return of more than 15% during the past decade. If you were to invest in only one ETF, this is a top choice given its long-term performance and diversity.

The Invesco QQQ Trust

Growth and tech stocks have helped lead the market higher during much of the past two decades, which makes the Invesco QQQ Trust (NASDAQ: QQQ) another great ETF option to dollar-cost average into over the long term. The ETF, which tracks the tech-heavy Nasdaq 100 Index, has consistently outperformed the S&P 500, topping it on a 12-month rolling basis 88% of the time during the past 10 years.

Although not as diversified as the S&P 500, the ETF has been a stellar performer, producing a 20.4% yearly return during the past decade. That's tough to beat, and with the technological innovation curve steepening, it could continue to outperform during the next decade.

Geoffrey Seiler has positions in Invesco QQQ Trust and Vanguard S&P 500 ETF. The Motley Fool has positions in and recommends Apple, Micron Technology, Nvidia, Palantir Technologies, Tesla, Vanguard S&P 500 ETF, and iShares Trust-iShares Semiconductor ETF. The Motley Fool has a disclosure policy.

The pace of artificial intelligence (AI) innovation continues to accelerate, and as it does, the need for faster connectivity is intensifying. This is leading to a shift in the data center away from traditional copper wiring to optical connections that are both faster and more energy-efficient. And rising demand for optical hardware is also creating what could be the next big bottleneck in the AI infrastructure build-out.

Let's look at eight AI stocks you could buy to play the optics trend, starting with chip companies and moving on to companies that provide the physical, campus, and wide-area networking (WAN) layers of optical infrastructure.

Optical DSPs

Marvell Technology (NASDAQ: MRVL) is the market leader in optical DSPs (digital signal processors), which serve as the brains inside high-speed optical transceivers. DSPs are the main chips used to take electrical signals from AI chip clusters and convert them into light, thus allowing the data to be transmitted through optical systems. Nvidia views Marvell's tech as so important to the data center tech stack that it made a $2 billion investment in the company and partnered with it to integrate Marvell's optical networking into its NVLink ecosystem.

Broadcom (NASDAQ: AVGO) is another big player in the optical DSP space, and its growth in this business is feeding off its large custom chip business.

Photonic lasers

Lumentum Holdings' (NASDAQ: LITE) laser tech is critical to preventing bandwidth bottlenecks in AI data centers. The company designs and manufactures high-power indium phosphide (InP) lasers, photodetectors, and optical engines that act as the raw light source for optical transceivers. The company is one of the few global players capable of producing these lasers at scale, giving it a virtual monopoly. Even when its competitors sell complete transceiver modules, they frequently use Lumentum's lasers to power them.

Transceivers

Coherent (NYSE: COHR) and Applied Optoelectronics (NASDAQ: AAOI) both provide critical transceiver infrastructure connecting switches and servers. Coherent is a powerhouse in optical materials, transceivers, and lasers, and its advanced photonics engine technology positions it as an essential partner to cloud providers. Since it is vertically integrated from the raw substrate material to the receivers themselves, the company has been one of the biggest beneficiaries of the shift toward using optical components in data centers.

Applied Optoelectronics is also a vertically integrated manufacturer of transceivers and components for data centers. It manufactures its own laser chips in-house for its transceivers, and has been aggressively ramping up manufacturing capacity to keep up with rising demand. The company is a key supplier to Microsoft, although this does add some customer concentration risk.

Image source: Getty Images.

Optical traffic

The final optics layer consists of the connectors, fiber, and systems that actually route data across server racks and between facilities. Amphenol (NYSE: APH) provides optical interconnect solutions that connect optical transceivers to switches inside a server rack, dramatically reducing latency and power consumption within GPU clusters. One great thing about the company, though, is that it will benefit from the data center boom regardless of how fast hyperscalers shift from copper to optics, since it not only supplies optical connectors, but also high-speed passive copper backplanes such as those used with Nvidia's NVLink interconnects.

Corning (NYSE: GLW) is another big beneficiary of the shift toward optical connectivity. Its high-density fiber cables and connection systems are the physical backbone of AI data centers, making this old-school glass company a next-generation optical play. Best of all, AI data centers require significantly more passive fiber strands than traditional data centers.

Finally, Ciena (NYSE: CIEN) builds optical routing systems that can carry optical traffic long distances from one data center to another. The stock is a nice way to invest in the increase of multifacility AI clusters and high-bandwidth optical transport.

Geoffrey Seiler has positions in Broadcom. The Motley Fool has positions in and recommends Amphenol, Broadcom, Ciena, Coherent, Corning, Lumentum, Marvell Technology, and Nvidia. The Motley Fool has a disclosure policy.

Stocks are often volatile around earnings, and even the slightest misstep can sometimes lead to big sell-offs. For long-term investors, though, these dips can be great buying opportunities, as the reasons behind them often have very little impact on a company's future prospects.

Sandisk (NASDAQ: SNDK), AppLovin (NASDAQ: APP), and Dutch Bros (NYSE: BROS) all crashed after earnings and now look like good long-term buys. Let's look at the case for each.

Image source: Getty Images.

Sandisk

If you were to look at Sandisk's recent fiscal fourth-quarter earnings in a vacuum, they were incredible. Its revenue surged 372% year over year to $9 billion, while its adjusted earnings per share (EPS) skyrocketed from $0.29 a year earlier to $39.25. The results were driven by soaring NAND (flash) memory prices, which drove revenue growth and helped its gross margin expand from 26.2% last year to 84.6%.

However, investors sent its shares sinking nearly 12% the following session as its fiscal Q1 guidance, which calls for revenue between $10.3 billion and $10.8 billion ($10.55 billion at the midpoint), came up just shy of the $10.62 billion consensus, and it projected its gross margin would slip slightly sequentially.

However, the big reason behind the "light" forecast was that Sandisk decided to forgo some near-term revenue and gross margin gains in favor of locking in longer-term five-year deals for more sustained growth. It now has eight contracts with revenue floor pricing of $93.9 billion and $16.5 billion in financial guarantees. This is actually the type of visibility investors should want to see from a company that has historically been in a very cyclical industry.

Trading at a forward price-to-earnings (P/E) ratio of 5.7, based on fiscal 2027 analyst estimates, the stock looks like a buy on the dip.

AppLovin

AppLovin is another company that saw robust revenue growth, but whose stock fell on high expectations. The company's revenue soared 53% to $1.92 billion, but that was just short of the $1.94 billion analyst consensus, sending its shares crashing 20% the next session.

The company said the revenue miss stemmed from its adtech AI model not improving at its usual speed, with the next big performance boost not coming until after the quarter ended. This led to a less robust pace of increased ad spending on its platform than expected, but it said the demand had already started to reaccelerate.

The plunge in the stock brought its forward P/E ratio to 16, based on 2027 analyst estimates, which is very cheap for a company projecting revenue growth of between 46% and 48% next quarter. This is a growth stock worth buying on the sell-off.

Dutch Bros

Dutch Bros shares sank nearly 17% after the coffee shop operator turned in another strong earnings report, as it forecast that its same-store sales growth would start to decelerate in the second half. However, its overall same-store sales growth remains strong and its expansion story remains unchanged.

In Q2, the company saw its revenue jump by 32.5% to $550.9 million, while EPS climbed 40% to $0.28. Its same-store sales rose by 5.8%, on a 1.7% bump in transactions, while company-owned comparable-store sales climbed 8.3% on a 3.4% increase in transactions. However, investors didn't like that Dutch Bros only raised the low end of its prior full-year same-store guidance, taking it from 4% to 6% to a new range of 5% to 6%.

Nonetheless, that is still solid same-store growth, and the company has a long growth runway of opening new stores. At the end of Q2, it had 1,225 stores, with plans to have over 2,000 by 2029 and a long-term target of 7,000 in the U.S. With impressive average unit volumes ($2.1 million), this is a company with tremendous growth ahead, making the stock a buy on the dip.

Geoffrey Seiler has positions in Dutch Bros. The Motley Fool has positions in and recommends Dutch Bros. The Motley Fool has a disclosure policy.

Hyperscalers, which are companies that own large data centers, have been spending aggressively to build out artificial intelligence (AI) infrastructure. Although the market has been concerned about this spending, the economics for things like cloud computing are quite good.

Amazon (NASDAQ: AMZN), for example, recently came out and said that it gets a payback on its AI chip and networking spending within two to three years, while SpaceX (NASDAQ: SPCX) has said it can get its money back within a year. Against that backdrop, let's look at one hyperscaler stock to buy, one to hold, and one to avoid.

Buy Amazon

Amazon is the world's largest cloud computing company, having invented the entire infrastructure-as-a-service concept. Today, it remains the market share leader and has been seeing accelerating growth, backed by partnerships with OpenAI and Anthropic, in which it holds about a 20% stake.

The company also has a strong proprietary custom-chip business, including its Trainium AI accelerators and Graviton central processing units (CPUs).

This is a fast-growing $25 billion annual recurring revenue (ARR) business that also helps it significantly reduce internal build costs and saves it money on inference. This is a big advantage that should become more important in the future.

Other parts of Amazon's business are often overlooked as well. It is the world's largest manufacturer and operator of robots and one of the largest digital advertisers. These two businesses are also helping drive strong operating leverage in its e-commerce segment.

Between its cloud computing growth, e-commerce strength, and large stake in Anthropic, Amazon stock is a buy.

Image source: Getty Images.

Hold Microsoft

Microsoft (NASDAQ: MSFT) has been one of the most beleaguered stocks in the cloud computing sphere amid concerns about its over its heavy reliance on OpenAI, lack of internal AI models and chips, and the fear that AI could undermine its core software business. However, after yet another strong earnings report, it looks like it may finally have brushed aside those issues.

Now, to be fair, the company still needs to catch up with its own AI, but it has started to make progress, and its investment and privileged partnership with OpenAI does give it time. The company still has huge commitments from OpenAI that should help power its cloud computing growth for years to come, while it has also shown that its software-as-a-service (SaaS) business remains sticky and is a primary way for enterprises to deploy AI.

Microsoft should continue to play a big role in AI over the long term, and it looks like a solid stock to continue to hold even after its big post-earnings run.

Avoid SpaceX

SpaceX is a lot of things, and cloud computing is becoming a bigger part of its story. It has a solid opportunity to grow in this area, while its ambitions to put data centers in space could become a game changer. However, it will have to overcome some major technical hurdles first, including protecting chips from cosmic radiation and developing a cooling system that works in the vacuum of space.

Meanwhile, the company's cloud unit appears to be overearning at this time. The company said it is getting paybacks as quickly as one year, but the reason behind this appears to be that it is getting strong pricing as a provider of bridge computing capacity. For example, Alphabet decided to unload some of its AI workloads to SpaceX at high prices rather than lose the business. This type of pricing is likely not sustainable over the long term, and with the company pledging to exclusively use Nvidia chips, it is at a cost disadvantage.

With SpaceX trading at a mind-blowing forward price-to-sales (P/S) ratio of 39 times 2026 analyst estimates, this is a stock I'd avoid. For reference, note that in its early days, Elon Musk's other company, Tesla, generally traded at a trailing P/S ratio of 10 to 12 times, so SpaceX is getting a huge valuation premium compared to his other stock.

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

Artificial intelligence (AI) capital expenditures are on the rise, and Morgan Stanley thinks the final number could come in even stronger than expected. The market is currently forecasting AI infrastructure spending to reach $1.2 trillion next year, but the investment firm thinks that figure may be too low and that it could hit $1.4 trillion in 2027.

Morgan Stanley analyst Erik Woodring pointed to recent commentary from the big four hyperscalers -- Amazon, Microsoft, Alphabet, and Meta Platforms -- with all four companies talking about industry capacity constraints. Meanwhile, the big cloud computing providers have mentioned that demand continues to outstrip their capacity and that they plan to significantly increase their capex next year.

Let's look at three AI stocks to benefit from this massive spending.

Image source: Getty Images

1. Nvidia

Nvidia (NASDAQ: NVDA) remains one of the best ways to play the AI infrastructure buildout. The company continues to see rapid growth, while the stock is cheap, trading at a forward P/E of 17 times fiscal 2028 (ending January 2028) analyst estimates.

The company has a dominant position in the AI model training market with its graphics processing units (GPUs), and its CUDA software platform, where most foundational AI code was written on, helps cement its leadership here. It has also nicely positioned itself for inference through its acquisition of Groq and its language processing units (LPUs), which help reduce latency during the important decode phase of inference. Meanwhile, the company also developed high-end central processing units (CPUs), which are becoming increasingly important with agentic AI.

As the market leader, and following a recent pledge from SpaceX to exclusively use its chips, Nvidia is a stock you want to own, as AI data center spending continues to surge.

2. Advanced Micro Devices

While the company isn't a big player in AI model training, Advanced Micro Devices (NASDAQ: AMD) looks poised to grab some meaningful share in the inference market, which is the faster-growing of the two markets.

Inference is very memory-intensive, and AMD's chiplet design can be packaged with more memory. It's also partnered with Cerebras, where its more expensive, but faster, systems can handle the de-code phase. In addition, AMD's recent acquisitions of memory optimization company MEXT and chip start-up Taalas, which boosts inference performance by embedding models directly into the chips, really show AMD's desire to become a major inference player.

At the same time, AMD is the leader in the data center CPU space. With the rise of inference and agentic AI, data centers will need a much narrower ratio of GPUs to CPUs to handle these tasks. AMD sees this becoming a $220 billion market over the next few years, with it continuing to be the market leader.

With AMD riding two of the most powerful trends in AI that are just getting started, and the introduction of its Helios rack system that combines its GPUs, CPUs, and networking into one system, this is a stock to buy.

3. Micron

One of the biggest bottlenecks in AI right now is memory, especially high-bandwidth memory (HBM), which gets packaged with GPUs and other AI chips to reduce latency and optimize performance. This becomes even more important with inference. Right now, demand for HBM is off the charts, and capacity additions are unable to keep up with increasing demand.

This has been great news for the big three memory makers, as surging DRAM (dynamic random access memory) prices have led to huge gains in revenue and gross margins. Micron (NASDAQ: MU), though, is arguably the biggest beneficiary of near-term increased AI spending, although much of this will be indirect.

The reason Micron benefits the most is that Samsung is still a conglomerate, and while SK Hynix is the HBM leader, regular DRAM prices have actually surged more than HBM prices because increasing HBM capacity is the primary focus of these three companies.

Expect the DRAM market to remain incredibly tight next year and well beyond, which should help lift Micron's stock. The stock is cheap with a forward P/E of below 6 times fiscal 2027 (ending August 2027) estimates, as investors wait for a cyclical shift that looks like it will keep getting pushed back.

Geoffrey Seiler has positions in Advanced Micro Devices, Alphabet, Amazon, and Meta Platforms. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Meta Platforms, Micron Technology, Microsoft, and Nvidia. The Motley Fool has a disclosure policy.

Inference has become the fastest-growing part of the artificial intelligence (AI) infrastructure market, and Bloomberg Intelligence projects it will double the size of the AI training market by 2032, reaching $1.3 trillion. With so much at stake, both leading chipmakers and upstarts are jockeying to grab a slice of this huge, fast-growing market.

Nvidia (NASDAQ: NVDA), Cerebras (NASDAQ: CBRS), and Advanced Micro Devices (NASDAQ: AMD) are all tackling this market in different ways. Let's see how these AI stocks stack up and why they could all be winners, given the size and growth of the inference market.

Image source: Getty Images.

1. Nvidia

Already the winner in AI model training, Nvidia now has its sights on the inference market. The company's big move to capture share was its "acquisition" of Groq and its language processing units (LPUs). Inference is more about fast memory access and low latency than raw compute power, and LPUs help address this by having SRAM (static random-access memory) embedded directly on their chips.

LPUs are particularly useful during the decode phase of inference, which is when large language models (LLMs) answer queries. As such, Nvidia now offers complete systems designed specifically for inference, where its graphics processing units (GPUs) handle the pre-fill phase (reading the prompt) while its LPUs handle the decode phase, thereby speeding up response times.

This is a nice solution and positions Nvidia to remain an AI infrastructure leader, even if it doesn't capture the same market share it does in training.

2. Cerebras

Like Nvidia, Cerebras is tackling inference with SRAM-based chips. However, because SRAM is so bulky, instead of just embedding a small amount onto its chips and stringing them together, Cerebras has created huge wafer-sized chips that are five to six times faster than LPUs.

The physical size of Cerebras' chips comes with some trade-offs. They require specialized cooling and energy management solutions and, as such, are only sold or rented as part of the Cerebras CS-3 systems. They also come at a very premium price tag.

However, the company has inked major deals with OpenAI and Amazon's AWS, and it recently announced a partnership with AMD that should help reduce the cost of ownership. The two companies will offer an inference solution in which AMD's Helios rack-scale solution will handle the pre-fill phase of inference, which it can do more cheaply, while Cerebras' Wafer-Scale Engine will perform the decode phase, which it can do more quickly. It's a nice way for companies to better compete with Nvidia's offerings.

Given the high cost of its systems, Cerebras has been more of a premium, niche solution, but it now looks set to become a major player in the humongous inference market.

3. AMD

After losing out on the LLM training market to Nvidia, AMD has been aggressively pursuing the inference market to make sure it doesn't get left behind again. Its chiplet design is better suited for inference, as it allows its GPUs to be packaged with more high-bandwidth memory (HBM) and to act as part of an entire unit to reduce latency. Meanwhile, its partnership with Cerebras looks like a smart move to help it better compete with Nvidia's complete inference system.

However, the company has not stopped there. It recently acquired memory optimization company MEXT and chip start-up Taalas to boost its inference offering. Memory is one of the biggest AI bottlenecks right now, and MEXT's solution can offload seldom-accessed data from DRAM to unused flash and then, using predictive AI, can transfer it back into DRAM before an application even requests it. This can reduce the need for more expensive DRAM and help save costs.

Meanwhile, Taalas has developed chips in which AI models are hardwired directly to bolster inference performance. Since the chips are model-specific, they aren't as flexible, but they are cheaper and much faster. The company plans to use them as part of a complete system where its GPUs would handle the pre-fill phase and Taalas' chips would handle the decode phase.

AMD is tackling inference from a couple of different angles, which should position it to capture a nice share of this huge market.

Geoffrey Seiler has positions in Advanced Micro Devices and Amazon. The Motley Fool has positions in and recommends Advanced Micro Devices, Amazon, and Nvidia. The Motley Fool has a disclosure policy.

Investor Warren Buffett may have retired as head of Berkshire Hathaway (NYSE: BRKA) (NYSE: BRKB), but his legacy and influence remain.

One of Buffett's main tenets when investing was to find great business models that could compound over decades. These types of businesses still make up the core of Berkshire's holdings and can be great long-term stocks to buy at any time.

Three of Buffett's favorite compounders include Apple (NASDAQ: AAPL), Coca-Cola (NYSE: KO), and Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG). Let's examine what makes each of these businesses so great.

Image source: The Motley Fool.

1. Apple: The Venus flytrap ecosystem

Apple is Berkshire Hathaway's largest holding, and for good reason: It arguably has the best compounding business model on the planet. The smartphone and PC industries typically follow a consistent replacement cycle, and Apple has carved out a strong niche at the high end of the market. Its products work seamlessly within a closed ecosystem, which attracts consumers for its ease and helps it command premium prices.

More importantly, this closed ecosystem eventually traps customers, as it is difficult to switch with each photo taken, app downloaded, and subscription purchased. This then feeds users into Apple's high-gross-margin services businesses. This includes cloud storage, commissions on app purchases to Apple Pay, and its revenue-sharing deal with Alphabet's Google Search.

It's this high-margin flywheel model, backed by an affluent customer base, that makes Apple a great stock to own long-term.

2. Coca-Cola: One of the world's most recognizable brands

Coca-Cola is one of Berkshire's oldest holdings, with Buffett buying the stock back in 1988 and never selling a single share. The reason the Oracle of Omaha loves Coca-Cola so much is that the company has unmatched brand equity and a great compounding business model.

While there are quite a few soda companies around the world, Coca-Cola has created an unmatched global moat through its advertising and distribution. Meanwhile, one key to the company's business is that it doesn't actually sell soda; it sells the syrup used to make its famous soda brands.

This shifts much of the heavy capital expenditures (capex) for owning the plants and delivery trucks onto its independent bottling partners, leaving it with a capex-light, high-margin business. It then feeds this into marketing and innovation, creating a flywheel effect that keeps its brands growing. With soda seeing a resurgence due to the increasing popularity of zero-sugar offerings, prebiotic alternatives, and dirty sodas (sodas mixed with flavored syrups and cream), now looks like a great time to own the stock.

3. Alphabet: The complete AI player

One of the last big purchases Buffett made before his retirement was Alphabet, another example of a great compounding business. This all starts with its core Search business, where it has built a huge distribution moat through its ownership of the web browser Chrome, smartphone operating system Android, and a search deal with Apple that essentially makes Google the gateway to the internet for most people.

Billions of daily search queries provide it with a continuous stream of behavioral data it can use to improve both search results and ad targeting. This, in turn, helps the company serve users more relevant ads, which leads to higher conversions and drives up cost per click. Once again, this is a great flywheel business.

Meanwhile, Google Cloud also has strong business economics, and its proprietary Tensor Processing Units (TPUs) enable Alphabet to run AI training and inference at a lower cost than competitors. This has led the business to achieve strong operating leverage and made the company one of the best-positioned companies for the future of AI, with its own top chips and AI models. This makes Alphabet a stock to own for the long haul.

Geoffrey Seiler has positions in Alphabet. The Motley Fool has positions in and recommends Alphabet, Apple, and Berkshire Hathaway. The Motley Fool has a disclosure policy.

It's a great feeling when a stock you own absolutely skyrockets higher after reporting earnings. Two software-as-a-service (SaaS) stocks, Atlassian (NASDAQ: TEAM) and Doximity (NYSE: DOCS), gave their investors something to cheer about when both companies' share prices skyrocketed more than 30% this past Friday, Aug. 7, following their earnings reports.

The two stocks had a few things in common going into the report. They both operate vertical-focused software platforms. In the case of Atlassian, its focus is on enterprise workflows, while Doximity's platform is centered on clinical workflows in the medical field. Both companies had also been widely viewed as potential losers in artificial intelligence (AI), and expectations heading into their earnings reports were pretty low. However, both stocks saw strong adoption of their AI offerings, which helped drive strong results and guidance.

With Atlassian and Doximity already seeing huge gains, let's look at three SaaS stocks that could have similar setups going into their upcoming earnings reports.

Image source: Getty Images

1. GitLab

While GitLab's (NASDAQ: GTLB) business is quite different from Atlassian's, the bear case for both has been quite similar. While Atlassian provides a platform to help enterprise teams collaborate and manage their work, GitLab runs a DevSecOps (development, security, and operations) platform that provides a secure ecosystem for organizations to design software. One of the big risks for both was that AI agents would lead to fewer seat licenses. As a result, both companies shifted toward hybrid consumption- and usage-based models and leaned into their own agentic AI tools.

This should actually be good for GitLab, as agents can generate more code, and it still needs to be written in a secure environment. The company has also been working on tools that make token consumption more efficient, which is something that's becoming highly sought-after as AI expenses rise. At the same time, commentary from Microsoft, which owns code repository GitHub, and JFrog, which plays a complementary role in the DevSecOps sector, could bode well for GitLab when it reports its Q3 results.

If GitLab's transition to a hybrid usage model shows early signs of driving revenue growth, the stock could be off to the races.

2. UiPath

Doximity was highly shorted going into its earnings report, with more than 15% short interest. UiPath's (NYSE: PATH) short interest is even higher, sitting around 25%. The stock is also very cheap, trading at a forward price-to-sales (P/S) ratio of 4 times 2027 analyst estimates and a forward P/E of 16.5. That's just a powder keg ready to explode if the company can deliver a strong report and guidance.

A leader in robotic process automation (RPA), the company's growth has slowed as organizations evaluate the use of software bots in an AI world. However, software bots can be better at deterministic, high-volume tasks, such as data entry or payroll, and they're certainly much cheaper. Meanwhile, UiPath's Maestro solution positions the company to be an agentic AI orchestration platform that can manage both AI agents and software bots with all the compliance and guardrail requirements. If Maestro can start to gain traction with the rise of AI agents and UiPath starts to see its annual recurring revenue growth accelerate, the stock has the potential to see a major pop come earnings time.

3. Asana

Another SaaS stock with low expectations is Asana (NYSE: ASAN). Like Atlassian, it operates a project management software platform, but instead of for IT departments, it is for non-technical teams such as marketing and human resources. The company has been beaten down over fears of slowing seat growth, the threat of AI disintermediation, and competition from Monday.com.

However, the company has been working to pivot away from smaller accounts to high-margin enterprise customers with better pricing. At the same time, it has cut costs and rolled out AI Studio (a no-code agent builder) and AI Teammates, which are autonomous agents that collaborate with entire teams, helping build shared organizational memory across workflows.

Notably, frontier model leader Anthropic has partnered closely with Asana, integrating its Claude AI model directly into AI Studio and AI Teammates to power its agentic reasoning. This partnership shows that foundation model companies like Anthropic view Asana more as a critical workflow layer rather than a target for disintermediation. If its AI add-on solutions can help accelerate revenue growth, the stock has the potential to rocket higher.

Geoffrey Seiler has positions in GitLab and UiPath. The Motley Fool has positions in and recommends Atlassian, Doximity, Microsoft, Monday.com, and UiPath. The Motley Fool recommends GitLab and JFrog. The Motley Fool has a disclosure policy.