Jensen Huang Tips Marvell as AI's Next Trillion-Dollar Giant — Stock Surges 25%

Jensen Huang, the chief executive of Nvidia, has identified Marvell Technology as the semiconductor company most likely to join the trillion-dollar club next — a prediction that sent the chip designer's shares surging 25% in after-hours trading on Thursday.

The endorsement came during a roundtable with analysts at Nvidia's Santa Clara headquarters, where Huang fielded questions about which companies stand to benefit most from the accelerating demand for AI infrastructure. When pressed on competitors worthy of long-term investment, Huang pointed to Marvell by name.

Marvell, based in Wilmington, Delaware, designs data center chips used in networking, storage, and AI acceleration. Unlike Nvidia, which manufactures its own GPUs, Marvell operates as a fabless designer — working with foundries like TSMC to produce its silicon. That business model has positioned the company to capture a growing slice of enterprise spending on AI hardware without the capital intensity of owning fabrication facilities.

The Case for a Trillion-Dollar Valuation

For Marvell to reach a trillion dollars in market capitalisation from its current level around $50 billion, the stock would need to climb roughly twentyfold from today's price. That sounds ambitious, but semiconductor analysts say the addressable market Marvell is pursuing has expanded dramatically over the past two years.

The company's custom AI accelerator business — chips designed specifically for individual hyperscale customers like Amazon Web Services, Microsoft Azure, and Google Cloud — has become a particularly valuable revenue stream. Unlike general-purpose GPUs, these custom silicon pieces are optimized for specific workloads, offering customers better performance-per-watt for tasks like AI inference and data processing.

Revenue from Marvell's AI-related businesses hit $2.7 billion in fiscal 2024, a figure the company projects will double by fiscal 2025. If that growth trajectory holds, Marvell could be generating $15 billion to $20 billion in annual AI revenue within five years — a scale that would justify premium valuations even by semiconductor standards.

Why Huang's Nod Matters More Than Analyst Reports

Industry endorsements carry weight in Silicon Valley, but few carry as much as Huang's. As the architect of Nvidia's dominance in AI training chips, Huang has unprecedented visibility into how the broader ecosystem is developing. When he speaks about which suppliers are winning business, institutional investors listen closely.

Marvell chief executive Matt Murphy has cultivated a relationship with major cloud providers over the past decade, positioning the company as a trusted partner for custom silicon projects. That strategy appears to be paying off at a moment when cloud giants are racing to develop their own AI chips to reduce dependence on Nvidia and control costs.

The endorsement arrives as Marvell prepares to ramp production of its 3-nanometer AI accelerators at TSMC's facilities in Taiwan. The company has committed to expanding its engineering headcount by 20% this year, with most of those roles based at its design centers in Austin, Texas, and Pune, India.

Market Reaction and Analyst Upgrades

Within two hours of the roundtable comments circulating on financial news wires, Marvell's share price had climbed from $58 to $72.50, adding roughly $12 billion to the company's market capitalisation in a single session. Trading volume exceeded 45 million shares — more than triple the stock's three-month daily average.

Several Wall Street firms revised their price targets upward following the news. Morgan Stanley raised its target from $65 to $85, citing Huang's comments as validation of Marvell's competitive positioning in custom AI silicon. Bank of America moved its target to $82, noting that hyperscaler demand for non-Nvidia AI chips appears to be accelerating faster than previously modeled.

Competition in the AI Chip Market Intensifies

Marvell's ascent, if it materializes, would unfold against a backdrop of intensifying competition in the semiconductor industry. Advanced Micro Devices, Intel, and dozens of startups are all competing for a piece of the AI infrastructure buildout estimated to exceed $300 billion annually by 2027.

The custom silicon segment — where Marvell competes directly with Nvidia's GH200 and H200 data center offerings — is particularly crowded. Amazon's Graviton processors, Google's Tensor Processing Units, and Meta's MTIA chips all represent efforts by cloud providers to bring more chip design in-house. That trend creates both opportunity and risk for Marvell: opportunity as a trusted design partner, risk if customers decide to bypass external vendors entirely.

Investors focused on the semiconductor supply chain note that Marvell's success hinges on TSMC's ability to ramp advanced node production. The Taiwanese foundry is operating near full capacity, and allocation disputes have become a persistent concern for fabless chip designers. Marvell has publicly stated it has secured sufficient wafer capacity through 2025, but longer-term commitments remain a subject of scrutiny.

What Comes Next for Marvell

The company is scheduled to report fiscal second-quarter earnings on August 29, and analysts will be watching closely for any updates on custom AI chip timelines. Marvell has previously guided that revenue from AI accelerators would reach $1.5 billion in the second half of fiscal 2025 — a number that will either validate or raise questions about the trajectory implied by Huang's comments.

Beyond earnings, investors should monitor Marvell's progress with its 5-nanometer and 3-nanometer roadmaps. The company has announced partnerships with two unnamed hyperscale customers for next-generation custom chips, with volume production expected to begin in the first half of 2026. The identity of those customers — and whether any are new logos — could provide further signal about Marvell's competitive standing.

The semiconductor sector's valuation multiples have compressed over the past six months as growth rates decelerated from pandemic-era peaks. Yet if Marvell's AI revenue continues to compound at 40% or more annually, the stock could revisit its 2021 highs above $90 within the next eighteen months, bringing the trillion-dollar horizon into sharper focus.