By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
Panelists agree that Marvell's interconnect business is a key driver for growth, but disagree on the impact of competitive pressures and hyperscaler vertical integration on the company's margins and market position.
Risk: Hyperscalers' vertical integration and potential in-house ASIC development could compress Marvell's margins and erode its market share.
Opportunity: Marvell's interconnect IP and CXL technology could provide platform lock-in and licensing revenue, securing its position in the AI memory fabric market.
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 →
Marvell turned in strong results and raised its guidance for both fiscal 2027 and fiscal 2028.
The company is seeing momentum driven by custom chips and optical interconnects.
Marvell Technology (NASDAQ: MRVL) has been one of the hottest stocks in the market this year, with its shares up 141% year to date, as of this writing. After significantly raising its revenue outlook for the next two years, the question is whether the semiconductor stock can continue its strong performance.
Let's dig into its fiscal first-quarter results and prospects to see if the stock is still a buy.
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 »
Marvell is riding two powerful trends with custom AI chips and optical data center components. It provides some important intellectual property (IP) for Amazon's Trainium chip and has also helped Microsoft develop its new Maia chip. It's also a leader in optical interconnects, providing critical networking connectivity for AI data centers.
In fiscal Q1, Marvell's fiscal Q4 revenue jumped 28% year over year to $2.42 billion. Data center revenue increased by 27% year over year in the quarter to $1.83 billion, while communication segment revenue climbed 29% to $585 million. Its adjusted earnings per share (EPS) rose 29% from $0.62 a year ago to $0.80. Those results were slightly ahead of the midpoint of management's outlook for adjusted EPS of $0.79 on revenue of $2.4 billion.
Looking ahead, Marvell is projecting fiscal Q2 revenue to grow 35% year over year to around $2.7 billion. It guided for adjusted EPS to be $0.93, up from $0.67 a year earlier.
For the full year, the company significantly upped its forecast, guiding for revenue to grow by 40% to nearly $11.5 billion. That was up from a prior outlook of $11 billion in sales, representing 30% growth. The growth is expected to be driven by its data center business, with revenue now anticipated to climb by 50%, up from an earlier 40% growth outlook. Within that segment, its interconnect business is forecast to surge by 70%, above its prior 50% projection.
This momentum is expected to continue in fiscal 2028, with Marvell projecting 45% revenue growth to $16.5 billion, above its prior $15 billion outlook. It expects its custom chip business to double.
Marvell's networking connectivity business, particularly interconnects, is seeing strong and increasing momentum, while its AI ASIC (application-specific integrated circuit) business is projected to see a huge lift next year. Meanwhile, the company's IP with SRAM (static random-access memory), which is the memory used in Nvidia's language processing unit and Cerebras' chips for inference, could become a nice growth driver down the line. However, there are some longer-term questions surrounding its Amazon chip business, with Taiwanese company AIchip reportedly fighting to capture more IP content.
The stock currently trades at a forward price-to-earnings (P/E) ratio of 37 times fiscal 2028 analyst estimates, which is a huge jump from earlier this year. Given the Amazon uncertainties, I'm not chasing the stock here, even though I do like its positioning in interconnects.
Before you buy stock in Marvell Technology, 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 Marvell Technology 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 $463,900! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,294,401!
Now, it’s worth noting Stock Advisor’s total average return is 978% — a market-crushing outperformance compared to 211% 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 31, 2026. *
Geoffrey Seiler has positions in Amazon. The Motley Fool has positions in and recommends Amazon, Marvell Technology, Microsoft, and Nvidia. 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
"Marvell's 37x FY2028 P/E already prices in flawless execution on interconnect and ASIC ramps that face credible competitive and customer-concentration risks."
Marvell's raised FY2027 revenue target of $11.5B (40% YoY) and FY2028 target of $16.5B (45% YoY), with data center now seen at 50% growth and interconnects at 70%, reflect real AI tailwinds in custom ASICs and optical components. Yet the stock's 141% YTD run and 37x forward P/E on FY2028 estimates leave little margin if Amazon's Trainium IP share erodes or if hyperscaler ASIC spending slows. The SRAM exposure to Nvidia/Cerebras is a secondary positive but unlikely to move the needle near-term. Execution risk on doubling custom chip revenue in FY2028 is underplayed.
If interconnect revenue actually compounds at 70% for two years while custom ASICs double, the 37x multiple could compress toward 25x on much higher earnings, making today's price look cheap rather than stretched.
"MRVL's interconnect business is genuinely strong, but the stock's 37x forward P/E leaves zero margin for error on Amazon competition or demand normalization."
MRVL's 141% YTD rally and 37x forward P/E on fiscal 2028 estimates screams valuation risk, not opportunity. Yes, 45% revenue growth through 2028 is real, but the article buries the critical issue: Amazon chip exposure faces competitive pressure from AIchip, and custom silicon is a lower-margin, higher-risk business than pure interconnect plays. The interconnect business (70% growth forecast) is the genuine moat, but it's already priced in. At 37x forward earnings, MRVL needs flawless execution and zero competitive surprises—a bet I wouldn't make at current levels.
If interconnects become THE bottleneck in AI infrastructure (not compute), MRVL's 70% growth could justify 40x+ multiples for years, especially if custom chips scale faster than the article assumes.
"Marvell's current valuation fully prices in aggressive growth, leaving the stock vulnerable to any minor deceleration in hyperscaler capital expenditure."
Marvell is effectively pivoting from a diversified semi-player to a pure-play AI infrastructure provider. While the 40% revenue growth guidance for FY25 is impressive, the valuation is the sticking point. Trading at 37x forward earnings for FY28 assumes near-perfect execution in the custom ASIC space. The real risk isn't just competition from AIchip; it's the cyclicality of the networking segment. If data center capex cools or hyperscalers like Amazon and Microsoft accelerate internalizing their silicon design cycles, Marvell's margins will face immediate compression. I am neutral because the current premium leaves zero margin for error in a sector prone to sudden inventory corrections.
The bull case rests on the 'moat' of their optical interconnect IP, which is becoming a standardized bottleneck in every AI cluster, potentially granting them pricing power that offsets any potential loss of custom ASIC market share.
"MRVL's continued hyperscaler AI capex and robust interconnect/IP monetization could justify a premium multiple and lift 2028 revenue toward $16.5B."
Marvell's FQ4 beat and guidance signal a durable AI data-center growth leg, with 2028 revenue seen at $16.5B (+45% y/y) and data-center/interconnect momentum (data center +50%, interconnect +70%). That aligns with AI infra buildouts and OEM IP wins for Trainium/Maia. The stock is pricey at ~37x fiscal 2028 estimates, so the upside hinges on a continued hyperscaler capex run and successful monetization of its custom chips. The Amazon IP angle remains a notable overhang; aIchip chatter hints at competitive risks. Also watch for margin pressure as you scale interconnects and potential AI-cycle slowdown. If the AI capex cycle cools or if key customers delay, MRVL could disappoint.
However, if hyperscalers throttle AI capex or adopt alternative interconnects, the revenue trajectory could stall, and the high multiple may compress faster than expected.
"Interconnect margin expansion could sustain the valuation if mix improves."
Claude underplays how interconnects could lift overall margins as their share grows. At 70% growth, optical components typically deliver 5-8 points higher gross margins than custom ASICs; if that mix reaches 35% of data-center revenue by FY2028, blended margins could expand rather than compress even with Amazon exposure. This offsets some AIchip risk the panel flags but remains unquantified in the discussion.
"Optical interconnects are becoming a commoditized bottleneck, not a durable moat—hyperscaler vertical integration threatens margin expansion, not just custom ASIC share."
Grok's margin math is sound but incomplete. Optical interconnects at 35% of data-center revenue by FY2028 would require that segment to grow faster than the stated 70%—it's currently ~15-20% of the mix. More critically: hyperscalers are vertically integrating optical switching too (see Microsoft's Myrtle). Margin expansion assumes MRVL retains pricing power in a commoditizing layer. The 5-8 point gross margin lift doesn't survive if customers build in-house.
"Marvell's long-term value lies in becoming the mandatory CXL-based interconnect standard for AI memory, which hyperscalers cannot easily replicate in-house."
Claude is right to flag vertical integration, but the panel is ignoring the 'CXL' (Compute Express Link) factor. Marvell isn't just selling optics; they are building the memory-pooling fabric essential for future GPU clusters. If CXL adoption accelerates, Marvell’s moat isn't just optical; it’s the architectural standard for AI memory scaling. This shifts the debate from 'margin compression' to 'platform lock-in.' Even if hyperscalers build their own ASICs, they will likely license Marvell's interconnect IP to ensure interoperability.
"CXL could create a platform moat via licensing, but could erode if hyperscalers push in-house memory fabrics, making the interconnect tail riskier than it sounds."
Responding to Gemini: the CXL angle is a double-edged sword. If CXL solidifies as the AI memory fabric standard, MRVL could gain platform-wide lock-in and licensing revenue, not just hardware sales. That’s a potential upside, but it also externalizes value, making MRVL reliant on ecosystem growth rather than chip margins. If hyperscalers push in-house memory fabrics and negotiate aggressively, the 'platform moat' could erode fast—making the 70% interconnect tail riskier than advertised.
Panelists agree that Marvell's interconnect business is a key driver for growth, but disagree on the impact of competitive pressures and hyperscaler vertical integration on the company's margins and market position.
Marvell's interconnect IP and CXL technology could provide platform lock-in and licensing revenue, securing its position in the AI memory fabric market.
Hyperscalers' vertical integration and potential in-house ASIC development could compress Marvell's margins and erode its market share.