By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel discusses AI infrastructure stocks MPWR, ALAB, and CDNS, highlighting their strong performances and growth potential, but also raises concerns about stretched valuations, cyclical risks, and potential threats from hyperscalers' custom silicon and capex digestion.
Risk: Potential cannibalization by hyperscalers moving to custom silicon and a sharp capex slowdown in 2026-27.
Opportunity: Potential for these enablers to capture more dollars per watt and per bit than the GPU vendor itself, even in a capex digestion phase.
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 →
Nvidia went on an incredible run to become the world's most valuable publicly traded company, but its 11% year-to-date return looks pedestrian compared to some of the other AI stocks that have been capturing headlines in recent months.
The three growth stocks on this list all have exposure to the AI infrastructure build-out, and they've all outgained Nvidia so far this year. They also look well positioned to extend their rallies and outperform it in 2027.
Where to invest $1,000 right now? Our analyst team just revealed what they believe are the 10 best stocks to buy right now, when you join Stock Advisor. See the stocks »
Monolithic Power Systems (NASDAQ: MPWR) produces power management systems that enable data centers to maintain continuous uptime without overloading AI chips. Power management systems work hand in hand with liquid-cooling solutions to keep chips and servers cool.
The stock has rallied by more than 40% year to date as AI data centers' demand for the company's products has grown. Monolithic Power Systems' revenue increased by 26.1% year over year in the first quarter, and net income grew at a slightly faster rate.
While the company lists six business segments in its earnings results, two of them are doing most of the heavy lifting. Enterprise data is the main one. It accounted for about one-third of total sales, and it nearly doubled year over year. This part of the business addresses power management and integrated solutions for AI chips and servers.
The communications segment is the other big one. This part of the business focuses on telecom infrastructure, satellite systems, and networking equipment. Its top line was up by 55.5% year over year, and up 33.1% sequentially, thanks to AI tailwinds. It makes up 14% of total sales.
As these two hypergrowth parts of the business gain market share, Monolithic Power Systems should experience accelerating revenue growth. That should position the stock to outperform Nvidia again in 2027.
Astera Labs (NASDAQ: ALAB) creates rack-scale connectivity hardware and software for AI servers. Many hyperscalers are turning to the company for connectivity solutions that enable faster data transmission between AI chips and server clusters.
Its revenue growth rates should prompt investors to give it a closer look. Sales almost doubled year over year in its first quarter, and its 14% sequential growth rate shows solid momentum. Double-digit percentage sequential revenue growth rates have become more common in the AI hardware space; for example, such a trend preceded Micron's incredible share price run.
Management's guidance is already pointing to meaningful growth from here. The $360 million midpoint of the guidance range for Q2 revenue implies 16.7% sequential growth. However, if recent history is any indicator, the actual growth rate may be closer to 20%. Astera Labs told investors to only expect up to $297 million in Q1 revenue, and yet it delivered $308.4 million.
Although the stock has almost doubled this year, it's also down by roughly 33% from its peak over the past few weeks, which presents a good buy-the-dip opportunity.
Cadence Design Systems (NASDAQ: CDNS) is off to a slower start than the other AI stocks on this list. It's up by almost 20% this year, driven by wins in electronic design automation software and hardware among chipmakers. These solutions enable companies like Nvidia and Advanced Micro Devices to test and build semiconductors before sending their designs to the third-party foundries that manufacture them.
Its growth rates are more moderate than the other two picks, but it sports a record $8 billion backlog. Cadence Design Systems also told investors to expect 17% revenue growth for 2026.
Agentic AI is set to be a major tailwind for the company. Its technology makes it easier to design advanced AI chips that can handle more rigorous workloads than those currently in data centers. That makes it a key checkpoint for AI chips, and that market can support revenue growth and profit margin expansion, which in turn would make the stock more attractive.
Before you buy stock in Monolithic Power Systems, 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 Monolithic Power Systems 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 $371,842! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,244,783!
Now, it’s worth noting Stock Advisor’s total average return is 900% — 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 July 18, 2026. *
Marc Guberti has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Advanced Micro Devices, Cadence Design Systems, Micron Technology, Monolithic Power Systems, and Nvidia. The Motley Fool recommends Astera Labs. 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
"While AI infrastructure demand is real, the article overstates 2027 outperformance odds by ignoring valuation resets, capex cycles, and NVDA's entrenched position."
The article highlights MPWR, ALAB, and CDNS as AI infrastructure plays that have already outpaced NVDA YTD (40%+, ~100%, ~20% respectively) with strong data-center tailwinds: MPWR's enterprise data/comms segments growing 2x and 55% YoY, ALAB's near-doubling revenue and 17% sequential guidance, and CDNS's $8B backlog plus 17% expected 2026 growth from agentic AI chip design. These are real secular drivers. However, valuations are stretched after the runs (ALAB especially volatile, -33% off highs), and the piece glosses over NVDA's ecosystem dominance, potential AI capex digestion in 2026-27, and cyclical risks in semis. Missing: MPWR's 26% revenue growth is solid but not explosive versus hyperscaler budgets; ALAB's beat-and-raise pattern isn't guaranteed forever.
The strongest case against is that these stocks' outperformance vs NVDA was largely multiple expansion on early AI hype; if 2027 capex slows or NVDA regains share in power/connectivity EDA, these names could easily lag or correct 30-50% while NVDA's scale and software moat keep it dominant.
"While infrastructure demand is durable, the current valuations for power and connectivity plays now bake in aggressive growth assumptions that leave zero margin for operational error."
The article correctly identifies that the AI trade is rotating from pure GPU compute to the 'picks and shovels' of infrastructure. Monolithic Power Systems (MPWR) and Astera Labs (ALAB) are essential for managing the thermal and data-bottleneck realities of H100/B200 clusters. However, the article ignores valuation risk. MPWR is trading at a significant premium, pricing in perfection in power delivery margins. Cadence (CDNS) is a defensive play on EDA (Electronic Design Automation) software, but its growth is tethered to semiconductor R&D cycles, which are notoriously lumpy. Investors should be wary of assuming these stocks will 'outperform' Nvidia simply because they have momentum; they are now facing the same law of large numbers that Nvidia is grappling with.
These companies are essentially high-beta derivatives of Nvidia's success; if Nvidia's demand cools due to hyperscaler capex fatigue, these suppliers will see their order books evaporate far faster than the primary chipmakers.
"These three stocks have genuine AI infrastructure exposure, but the article provides no valuation framework, competitive analysis, or downside scenarios—making it impossible to assess whether outperformance is justified or just momentum."
The article conflates outperformance with future outperformance—a classic recency bias trap. MPWR, ALAB, and CDNS have indeed beaten NVDA's 11% YTD return, but that's partly because NVDA started from a higher base and faced profit-taking. More critically: MPWR's 26% revenue growth is solid but not exceptional for AI infrastructure; ALAB's 14% sequential growth, while impressive, sits atop a $360M revenue base—scale matters for sustainability; CDNS at 20% YTD is the slowest of the three, and the 'agentic AI' thesis is speculative. The article also omits valuation entirely. Without forward P/E multiples, margin profiles, and competitive moats, 'outperform Nvidia' is marketing, not analysis. These are real companies in real growth markets, but the article reads like a momentum play dressed as thesis.
If NVDA re-rates downward due to margin compression or slowing capex cycles, these infrastructure plays could get dragged down harder—they're leveraged bets on the same hyperscaler spending, not hedges against it.
"Nvidia's platform leadership and scale make it unlikely these peers will consistently outpace NVDA through 2027."
The Motley Fool piece casts MPWR, ALAB, and CDNS as AI infra winners and suggests they’ll outpace Nvidia in 2027. But the bulls’ thesis hinges on a multi-year, broad-based AI capex cycle and ongoing tailwinds that may already be priced in. The strongest counterpoint is Nvidia’s enduring platform moat—CUDA software, ecosystem, and scale in AI training/inference—which supports pricing power and continued data-center demand even if hardware suppliers face cyclicality. Astera Labs and Cadence are exposed to capex timing and project-driven budgets, which can swing quarterly and annual results. A sustained AI-cycle slowdown or a turn in capex could compress multiples for these names even as Nvidia remains resilient.
Nvidia's moat could widen further and keep leading the cycle, while MPWR, ALAB, and CDNS are highly cyclical and prone to margin compression if AI capex eases. Therefore, the outperformance thesis may be fragile rather than durable.
"Infrastructure names can still outperform on content-per-rack gains even if total AI capex plateaus."
Claude correctly flags valuation omission, but all four underplay second-order risk: if hyperscalers hit 2026 digestion phase, MPWR/ALAB's power and connectivity content per rack could still grow 30-40% while NVDA's GPU unit growth flattens. The real divergence isn't capex slowdown—it's whether these enablers capture more dollars per watt and per bit than the GPU vendor itself.
"Hyperscaler vertical integration into custom silicon poses an existential margin risk to component suppliers like MPWR and ALAB that NVDA’s software moat is better insulated against."
Grok, you're missing the 'moat' reality: MPWR and ALAB are commodity-adjacent. If hyperscalers like Amazon or Google move to custom silicon (ASICs), they will prioritize internal power/connectivity solutions, directly cannibalizing MPWR and ALAB's margins. NVDA’s CUDA moat is sticky, but hardware suppliers are vulnerable to the 'white-label' shift. We are ignoring the risk that hyperscalers aren't just digesting capex—they are actively looking to bypass these specific vendors to improve their own unit economics.
"Hyperscaler ASIC shifts are real but slow; margin pressure from competitive bidding is the nearer-term threat than outright displacement."
Gemini's ASIC cannibalization risk is real, but underestimates switching costs. Custom power delivery and connectivity silicon take 18-24 months to design and validate—hyperscalers can't pivot overnight. MPWR/ALAB have time to lock in design wins and contracts. The bigger threat isn't ASIC replacement; it's margin compression if these suppliers compete on price to prevent it. That's a 2027-28 problem, not immediate.
"The real risk to MPWR/ALAB is a cyclical capex downturn that compresses volumes and margins, not just hyperscaler cannibalization."
Gemini, your ASIC-cannibalization risk matters, but timing matters more. Even with internal silicon, hyperscalers still require MPWR/ALAB for reliable power delivery and high-speed interconnects—design wins stick, switching costs are nontrivial. The bigger risk is a sharp 2026-27 capex slowdown that pressures volumes and margins across suppliers, not just cannibalization. The thesis hinges on durable wallet share and cyclical demand resilience, not a one-way shift to in-house silicon.
The panel discusses AI infrastructure stocks MPWR, ALAB, and CDNS, highlighting their strong performances and growth potential, but also raises concerns about stretched valuations, cyclical risks, and potential threats from hyperscalers' custom silicon and capex digestion.
Potential for these enablers to capture more dollars per watt and per bit than the GPU vendor itself, even in a capex digestion phase.
Potential cannibalization by hyperscalers moving to custom silicon and a sharp capex slowdown in 2026-27.