By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panel consensus is bearish on Marvell Technology (MRVL) due to execution and concentration risks in its XPU ramp, margin compression concerns, and uncertainty around ASP increases and cross-ecosystem monetization. While Marvell's optical interconnect business shows strong growth, the panel questions whether a single $10B customer can justify a $1T valuation.
Risk: Concentration risk in the XPU ramp and potential episodic nature of the Microsoft/XPU relationship
Opportunity: Potential shift to high-margin licensing if Marvell successfully commoditizes the custom ASIC space
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 →
Huang said this company makes chips that are "essential" for the future of artificial intelligence (AI).
Nvidia partnered with the chipmaker earlier this year to ensure compatibility with its NVLink platform.
The chipmaker's accelerating revenue growth should continue for years.
Nvidia (NASDAQ: NVDA) CEO Jensen Huang is one of the most well-respected names in the artificial intelligence (AI) industry. As the head of the largest AI semiconductor company, working closely with the world's leading cloud computing companies, he has unique insight into their evolving needs.
That's why investors pay close attention to what Huang has to say. On stage at Computex Week in Taipei, Huang presented with Marvell Technology (NASDAQ: MRVL) CEO Matt Murphy, describing his company's chips as essential. "That's why you're going to be the next trillion-dollar company," Huang said.
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 stock has skyrocketed since those comments, and it sports a market cap of about $278 billion as of this writing. If Huang is right, though, the stock could nearly quadruple from here.
Marvell specializes in optical interconnect chips that transmit data at ultra-high speeds. As Huang said, those chips are essential for optimizing data centers full of expensive, high-powered GPUs. Ensuring data is routed to where it needs to be as quickly as possible can provide an enormous cost savings at the hyperscale level.
To that end, Marvell and Nvidia agreed earlier this year to work closely, using Nvidia's NVLink platform to optimize chip designs across Marvell's portfolio. That coincided with Nvidia investing $2 billion of its cash in Marvell stock.
Management raised its outlook for the interconnect business with its first-quarter earnings report last month. It now expects 70% year-over-year growth for the segment. That should support overall revenue growth of 40% for the full year, with continued acceleration into 2028.
Another key driver of revenue growth in fiscal 2028 and beyond is Marvell's custom AI accelerator (XPU) business. Management said it expects the relatively small segment to grow 20% this year, but it will double in 2028 as a major tier-1 XPU customer (likely Microsoft) begins volume production. Management forecast $10 billion in revenue from custom AI chips by fiscal 2029. To put that in perspective, that's more than Marvell's total revenue last year.
There's no doubt Marvell is well positioned to generate substantial revenue growth over the next few years, thanks to its lead in networking chips and rapidly growing custom compute business. While Huang didn't set a timeline for his $1 trillion declaration, Marvell would need to reach a trailing price-to-sales ratio of about 40 in early 2029, based on management's guidance and commentary. Its trailing P/E ratio would have to reach 127 based on analysts' current estimates. For reference, Nvidia currently trades for 21 times sales and 33 times earnings.
For Marvell to reach a $1 trillion valuation within the next few years, it'll have to exceed management's own guidance and analysts' expectations by a wide margin. That said, Marvell is an excellent business with strong momentum in two key areas of artificial intelligence semiconductors. Patient investors may be rewarded for holding the stock. However, after the recent run-up in the stock price driven by Huang's comments, readers may want to wait for a better entry point.
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 $443,191! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,258,838!
Now, it’s worth noting Stock Advisor’s total average return is 941% — 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 June 6, 2026. *
Adam Levy has positions in Microsoft. The Motley Fool has positions in and recommends 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 upside relies on an extraordinary, uncertain AI-driven growth path that is not yet supported by current fundamentals or credible cross-customer demand."
Today’s piece leans into Jensen Huang’s hype, portraying Marvell as Nvidia’s essential interconnect partner and a potential 'next trillion-dollar' AI chip supplier. The thrust hinges on NVLink-enabled data-center efficiency, a booming XPU roadmap, and a Microsoft-backed growth surge. But the whole argument rests on aggressive projections: 70% interconnect growth, 40% full-year revenue growth, and $10B in AI-chip revenue by 2029. Those targets imply massive margin expansion and market-share gains that are not yet supported by current fundamentals. It's a narrative that depends on a durable AI capex cycle and a single ecosystem win, which is a risky bet.
Bull case against my stance: if AI capex stays structurally elevated and Marvell wins a Microsoft-scale XPU deal, the interconnect ramp could materialize faster than expected, justifying higher valuation.
"Marvell’s path to a $1 trillion valuation requires a multiple expansion that ignores the high probability of margin compression as custom AI silicon becomes a commoditized utility."
Marvell Technology (MRVL) is riding the coattails of Nvidia’s (NVDA) infrastructure dominance, but the $1 trillion valuation target is pure hyperbole. While the 70% growth in optical interconnects is impressive, it relies on the assumption that hyperscalers won't eventually bring this silicon design in-house to capture margin. Marvell’s shift toward custom AI accelerators (XPUs) is a high-stakes pivot; if they lose a major tier-1 customer or if the custom silicon market commoditizes, the 40x price-to-sales multiple required for a trillion-dollar cap becomes a fantasy. Investors are currently pricing in perfection, ignoring the cyclical nature of semiconductor capital expenditures and the inevitable margin compression that follows custom ASIC scaling.
If Marvell successfully becomes the 'Intel of the AI era' for networking, their proprietary IP could create a moat so deep that hyperscalers are forced to pay a permanent premium, justifying the valuation expansion.
"Marvell needs to crush its own $10B custom-AI forecast and sustain 40x sales multiples to justify a trillion-dollar valuation within five years."
The Nvidia endorsement and NVLink tie-up validate Marvell's interconnect edge, while the custom XPU ramp (targeting $10B by FY2029, likely via Microsoft) offers a credible second growth vector beyond the 70% interconnect surge. Yet the $278B market cap already prices in aggressive 40%+ revenue growth through 2028; reaching $1T would demand a 40x sales multiple by 2029 versus Nvidia's 21x, implying either massive guidance beats or multiple expansion that history rarely grants networking semis. Customer concentration and execution risk on the XPU side remain under-discussed.
Marvell could still compound faster than modeled if hyperscaler custom-silicon demand accelerates beyond the single tier-1 customer, pushing revenue and margins well above the $10B FY2029 target.
"Marvell's interconnect business is solid, but reaching $1T requires XPU execution that depends on a single customer and valuation multiples 2x Nvidia's—current pricing has already baked in the upside."
Marvell's optical interconnect business is genuinely valuable—70% YoY growth in a $278B market cap is real momentum. But the article conflates Huang's throwaway compliment with a valuation target. To hit $1T by 2029, Marvell needs a 40x P/S ratio; Nvidia trades 21x. The XPU business ($10B by 2029) is speculative—dependent on a single unnamed customer (likely Microsoft) ramping volume production. That's concentration risk the article barely mentions. Huang's comment was marketing theater, not a forecast. The stock has already repriced on the partnership and Nvidia's $2B investment.
Marvell could genuinely become a $1T company if AI infrastructure spending accelerates beyond consensus and the XPU business achieves even 60% of management's $10B target—the optionality alone justifies current valuations for growth-focused investors with long time horizons.
"The '1T by 2029' path relies on two uncertain bets—durable XPU demand from a single customer and outsized margin/valuation expansion—that may not materialize; without them, the stock’s premium is fragile."
Gemini overstates the valuation risk; my concern is execution and concentration risk in the XPU ramp. Even with 70% interconnect growth, MRVL faces margin compression if hyperscalers push in-house designs or if the Microsoft/XPU relationship proves episodic rather than durable, triggering downside to revenue and free cash flow. The 1T thesis hinges on two uncertain bets: big increases in ASPs and sustained cross-ecosystem monetization, which history hasn't rewarded in midcaps.
"Marvell’s potential to shift from hardware vendor to high-margin IP licensor justifies a valuation premium that purely hardware-focused models miss."
Claude, you correctly identify the concentration risk, but you overlook the structural shift in networking. It isn't just about 'marketing theater' from Huang; it’s about the shift from Ethernet to proprietary interconnects where Marvell holds the IP moat. While the 40x P/S target is absurd, the market is pricing in a 'platform' premium, not just a component supplier. If they successfully commoditize the custom ASIC space, their margin profile shifts from hardware to high-margin licensing, fundamentally changing the valuation ceiling.
"Marvell's interconnect edge stays Nvidia-tethered and cannot independently create a licensing moat."
Gemini claims Marvell's interconnect IP will force a permanent platform premium and shift to licensing margins, yet this overlooks how NVLink remains Nvidia-controlled while Marvell supplies components. Any Microsoft XPU win stays captive to one customer's roadmap and ASP negotiations. Without multiple tier-1 design wins, the supposed moat collapses into standard semi cyclicality rather than structural re-rating.
"Marvell's interconnect moat is real but geographically and customer-bound, not platform-wide—one Microsoft win doesn't justify trillion-dollar optionality without evidence of a second tier-1 design win by mid-2027."
Grok's point about NVLink remaining Nvidia-controlled is crucial, but it undersells Marvell's actual leverage: they own the optical interconnect layer *between* Nvidia GPUs and custom ASICs. That's different from being a component supplier. However, Grok is right that without multiple tier-1 wins beyond Microsoft, the moat remains customer-specific rather than structural. The real question: does one $10B customer justify a $1T valuation, or does it require proof of repeatability by 2027?
The panel consensus is bearish on Marvell Technology (MRVL) due to execution and concentration risks in its XPU ramp, margin compression concerns, and uncertainty around ASP increases and cross-ecosystem monetization. While Marvell's optical interconnect business shows strong growth, the panel questions whether a single $10B customer can justify a $1T valuation.
Potential shift to high-margin licensing if Marvell successfully commoditizes the custom ASIC space
Concentration risk in the XPU ramp and potential episodic nature of the Microsoft/XPU relationship