By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel agrees that Astera Labs (ALAB) is overvalued at 45x forward earnings, with risks including cyclicality, customer concentration, and potential margin compression due to competition from Broadcom and Marvell. However, there's disagreement on whether ALAB's PCIe connectivity provides a lasting moat.
Risk: Customer concentration and potential margin compression due to competition and commoditization of interconnects.
Opportunity: ALAB's specialized role in data center interconnects and the potential for maintaining pricing power in the near term.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
You know it first and foremost as a personal computer company, but Dell is making a big splash in the artificial intelligence platform space too.
To make the most of AI’s capabilities, the world needs a means of gathering physical data and then doing something mechanical with it. ON Semiconductor has solutions for both.
Astera Labs solves problems that AI data center developers didn’t initially realize they’d be facing.
You're certainly familiar with names like Nvidia and Palantir Technologies. The former remains the world's chief supplier of artificial intelligence (AI) data center processors, while the latter is one of the most-used decision-intelligence platforms. Both stocks have performed very well of late thanks to AI mania.
The problem with stepping into such well-known names, however, is simply that these trades can be very crowded and therefore very expensive. As Warren Buffett famously advises, "You can't buy what is popular and do well." Oh, there are clear exceptions to his argument -- both Palantir and Nvidia continued rallying well after both tickers became well-known must-haves. Plenty of investors understandably suspect that these two stocks' highest-growth phase is in the rearview mirror. Smart investors are looking for the next unknown AI gem that's yet to be discovered and subsequently fully valued.
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
To this end, three under-the-radar AI stocks to buy this year before their underlying companies become a more important part of the AI conversation are Dell (NYSE: DELL), ON Semiconductor (NASDAQ: ON), and Astera Labs (NASDAQ: ALAB). Here's why.
Dell is not only still around as a major brand in the desktop and laptop space, but is also inching into the AI space with a platform called Dell AI Factory.
In simplest terms, Dell AI Factory allows organizations to harness the power of artificial intelligence in a way that's easy to implement and easy to use... using Dell's hardware of course (although often paired with Nvidia-made processors). Formula 1 racing team McClaren, energy and chemical outfit Worley, and retailer Lowe's are just some of the organizations that have been able to do something constructive with otherwise abstract and difficult-to-build AI tech.
And it's working! Last fiscal year's top line grew 19% to a record-breaking $113.5 billion, led by 40% growth from its infrastructure solutions group, which was led by an explosion of sales of its AI-optimized servers. Analysts are looking for comparable top-line and bottom-line growth this year as well. Its practical, turnkey offerings are the option that many companies not interested in piecing together their own AI solution have been waiting for.
Just be careful if you're interested. While most investors don't yet think of Dell as a participant in the AI revolution, enough of them have found and plowed into a stake in this $160 billion company to push shares up to a price that's 30% above Wall Street's consensus target of $191.21.
Much of this gain has occurred only recently, so analysts may not have had a chance to update their stances. Still, if you can hold out for a slightly better entry point, Dell is one of the stock market's better-kept AI secrets.
Creating a powerful AI platform is one thing. Doing something constructive with it is another. AI still needs a way to convert physical information into digital data and then do something mechanically useful with its computed information.
ON Semiconductor is quietly bridging that gap.
Simply put, ON makes a range of industrial sensors, wireless antennas, and microcontrollers, along with power controllers, high-capacity semiconductors, and motor controllers that are used in everything from driver-assistance tech to medical diagnostic equipment to factories to wearables, and more. The company's current developmental partners include electric vehicle (EV) makers Geely and Nio, but it's also working with China's Sineng Electric on energy-storage solutions. It's even partnered with Nvidia to develop new 800-volt power solutions that the next generation of AI data centers is likely to utilize to improve power efficiency.
It's not a high-growth business yet, for the record; double-digit revenue growth is still a very good year for this company. Its revenue and earnings growth are apt to accelerate in the foreseeable future, however, driven by its soup-to-nuts offerings at a time when factories, automobiles, healthcare, and even cities are becoming more AI-automated.
ON's a consistent grower in the meantime and usually profitable. And when it isn't, it's often for non-operational reasons like last quarter's restructuring impairment charge. This fiscal viability makes it something of a standout compared to many of its direct competitors like Navitas and Wolfspeed.
Last but not least, add Astera Labs to your list of overlooked or unknown AI stocks to buy in 2026 before the crowd discovers its potential.
In the industry's infancy, AI data centers were built using existing, off-the-shelf components like Nvidia's graphics processing units (GPUs), networking hardware from Cisco, and PC memory chips from Micron. And it was fine... in the beginning. It didn't take the business very long to realize it was consuming and creating more digital information than this generation of equipment could handle. It needed more, but it also needed cost-effective solutions capable of integrating older hardware with newer components.
Astera Labs answered the call.
In simplest terms, Astera designs and manufactures entire systems that interconnect an AI data center's thousands of processors. Specifically, its Aries retimers and cables receive and deliver high-speed signals from processors, its Scorpio fabric switches make the most of available bandwidth, its Leo memory controllers improve the existing memory capacity of legacy physical interfaces, while its Taurus ethernet cards dramatically improve traditional networking solutions. Astera Labs also offers the software that makes all of this hardware work together to achieve some pretty amazing optimization. That's why its list of customers and partners consists of hyperscalers like Microsoft and Amazon.
The company's time is finally here. Last fiscal quarter's revenue of $308.4 million was 14% better than the previous quarter's and 93% higher year over year. Analysts expect comparable revenue growth this year and next to drive even more explosive earnings growth. Yes, Astera Labs is profitable too, on pace to report nearly $3 per-share profit in 2026, en-route to an expected $4.33 for 2027.
It's not a cheap stock, priced at 45 times next year's projected earnings. That's not terribly expensive, however, given the long-term opportunity at hand. Industry research outfit Global Market Insights expects the worldwide data center infrastructure market that Astera serves to grow at an average annual pace of 13.4% through 2034.
Before you buy stock in Astera Labs, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Astera Labs wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $471,827! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,319,291!
Now, it’s worth noting Stock Advisor’s total average return is 986% — a market-crushing outperformance compared to 207% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of May 10, 2026. *
James Brumley has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Amazon, Cisco Systems, Micron Technology, Microsoft, Nvidia, and Palantir Technologies. The Motley Fool recommends Astera Labs, Lowe's Companies, ON Semiconductor, and Wolfspeed. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"These picks are not 'under-the-radar' gems but rather high-beta infrastructure plays that are already priced for perfection in a cooling hardware cycle."
The enthusiasm for Dell, ON, and Astera Labs overlooks the cyclical nature of hardware and the extreme valuation premiums already baked in. While Dell’s 'AI Factory' pivot is impressive, it is essentially a high-volume, low-margin assembly business that remains tethered to Nvidia’s supply chain constraints. Astera Labs (ALAB) is the most interesting play here due to its specialized role in data center interconnects, but trading at 45x forward earnings leaves zero room for execution errors or a slowdown in hyperscaler CapEx. ON Semiconductor is the most speculative; its reliance on the EV sector, which is currently facing significant margin compression, makes it a 'value trap' rather than a pure AI play.
The bull case for these names rests on the assumption that AI infrastructure spending is a secular multi-year cycle rather than a temporary capital expenditure bubble that will inevitably face a 'digestion period' once capacity meets current demand.
"ALAB's 45x fwd P/E leaves no margin for error if AI infrastructure growth moderates to the 13% market CAGR cited."
This Motley Fool piece hypes DELL, ON, and ALAB as overlooked AI plays, but glosses over stretched valuations: DELL trades 30% above $191 consensus targets amid recent surges, while ALAB's 45x 2026 fwd P/E (price-to-earnings multiple on projected earnings) embeds explosive growth that Q1 hyperscaler guidance could deflate. ON's AI bridge via sensors/power tech is real but dwarfed by cyclical auto/EV exposure (e.g., Geely/Nio partners), vulnerable to demand slowdowns. Missing context: AI capex boom risks peaking as Microsoft/Amazon optimize racks, per recent earnings; commoditization threatens pick-and-shovel margins long-term.
If AI data center buildout sustains 20%+ CAGR through 2030 amid agentic AI breakthroughs, these stocks' 'under-radar' status could fuel Nvidia-like re-ratings from current levels.
"These stocks are priced as if they're hidden gems but are already re-rated; the real risk is that AI infrastructure capex cycles are front-loaded and competitive moats are thinner than the article suggests."
This article conflates 'under-the-radar' with 'already re-rated.' DELL trades 30% above consensus; ALAB at 45x forward P/E isn't hidden—it's priced for perfection. The real tension: these are *enablers* of AI infrastructure, not AI itself. They benefit from capex cycles, but those cycles are lumpy and competitive. ON Semiconductor's 'consistent grower' framing masks cyclicality in semis and automotive exposure. The article cherry-picks customer wins (McLaren, Lowe's) without addressing TAM saturation risk or margin compression as competition intensifies. Most critically: if Nvidia's highest-growth phase is 'in the rearview,' why assume these suppliers' isn't?
If AI capex inflection is real and sustained through 2027–2028, these picks could be legitimately cheap relative to their addressable market, and the article's 'not yet discovered' framing may simply be correct—early institutional adoption hasn't priced in enterprise adoption.
"If AI data-center capex stays resilient, Dell, ON Semiconductor, and Astera Labs can compound despite rich valuations; if spending slows, the thesis faces meaningful multiple compression."
This piece pushes a tilt toward under-the-radar AI hardware names—DELL, ON Semiconductor, and Astera Labs—as a way to play AI without chasing Nvidia/Palantir. Dell AI Factory augments a massive PC/bookending hardware business, but AI revenue remains a slice and the stock trades at a premium to peers. ON Semiconductor offers broad AI-enabled automation exposure, yet it’s cyclical and tied to factory and EV capex. Astera Labs trades near 45x forward earnings with a long-run data-center market CAGR around 13%, a high bar if AI demand softens or supply dynamics normalize. If data-center capex remains robust, the thesis could compound; otherwise, multiple compression looms.
The strongest counter is that AI hardware spend may decelerate faster than anticipated, making these names look rich on the back of hype rather than durable fundamentals.
"Astera Labs possesses a structural moat as a critical bottleneck provider, justifying a premium valuation despite broader hardware cyclicality risks."
Claude, you’re right that these are enablers, but you’re ignoring the 'bottleneck' premium. If hyperscalers prioritize interconnect efficiency to maximize GPU utilization, Astera Labs isn't just a supplier; they are a tax on every high-end data center build. While others fear margin compression, I see a moat in PCIe connectivity that isn't commoditized yet. The risk isn't just cyclicality—it's whether these firms can maintain pricing power once custom silicon starts replacing merchant hardware.
"ALAB's interconnect moat faces rapid erosion from competitors and high customer concentration risks."
Gemini, ALAB's 'bottleneck premium' in PCIe ignores accelerating competition from Broadcom's Jericho3-AI and Marvell's custom ASICs, per their Q1 earnings—standards like CXL 3.0 commoditize interconnects faster than you imply. Customer concentration (top client ~45% of revenue, SEC filings) means one hyperscaler pivot crushes margins. This isn't a moat; it's a timing bet on delayed alternatives.
"Standardization risk is real but timeline matters—Astera has 18–24 months to defend before CXL 3.0 commoditizes, and the market isn't pricing that binary outcome."
Grok's competitive threat is real, but conflates standardization with margin collapse. CXL 3.0 adoption takes 3–5 years; Astera's revenue concentration is a *near-term* risk, not proof the moat evaporates. The harder question: does Astera pivot to custom silicon before Broadcom/Marvell scale, or does it become a niche player? That's a 2026–2027 inflection, not priced into current 45x multiple.
"ALAB's 45% revenue concentration makes it vulnerable to a single hyperscaler renegotiation or capex pullback, undermining any moat from PCIe bottlenecks and risking margin compression if AI capex softens."
Grok, you frame ALAB’s moat as a premium from PCIe bottlenecks, but the obvious counter is customer concentration: ~45% of revenue from a single hyperscaler means one renegotiation or capex pullback can disproportionately crater margins. Even if Jericho3-AI and Marvell scale, the path to durable pricing power looks hinge-laden, not moat-protected, and 3–5 years of CXL standardization may not save profits if capex cools.
The panel agrees that Astera Labs (ALAB) is overvalued at 45x forward earnings, with risks including cyclicality, customer concentration, and potential margin compression due to competition from Broadcom and Marvell. However, there's disagreement on whether ALAB's PCIe connectivity provides a lasting moat.
ALAB's specialized role in data center interconnects and the potential for maintaining pricing power in the near term.
Customer concentration and potential margin compression due to competition and commoditization of interconnects.