Nvidia's CEO Jensen Huang Just Called Marvell Technology The Next Trillion-Dollar Company. But Is It the Better Stock to Buy?

Motley Fool highlights Nvidia’s stronger growth versus Marvell: analysts expect Marvell revenue growth of 41% this year and 45% next, versus Nvidia’s 81% this year and 41% next. Valuation is also cited as a key advantage, with NVDA trading at ~22x forward earnings versus MRVL near ~70x, implying Marvell must more than triple earnings to justify similar valuation. Despite Jensen Huang’s praise for Marvell as a potential “next trillion-dollar” company, the article ultimately favors Nvidia as the better risk/reward AI stock to buy and hold.

Analysis

The market is still underpricing how asymmetric AI monetization is between a platform vendor and a custom-silicon hopeful. NVDA’s moat is not just compute share; it is the default liquidity source for hyperscaler AI capex, which means any delay in in-house ASIC adoption tends to re-route spend back to GPUs rather than disappear. MRVL’s upside is real, but it is a second-order claim on future design wins, and that makes it much more execution-sensitive than the narrative implies.

The key second-order effect is that MSFT and AMZN pushing custom silicon does not necessarily “take share” from NVDA on a one-for-one basis in the next 2-4 quarters. It more likely shifts budget toward networking, interconnect, and systems integration while keeping a large bridge demand for merchant accelerators; that favors the broader AI infrastructure complex, but not equally. In that setup, MRVL’s valuation is the fragile point: if its revenue inflects slower than the market expects, multiple compression can dominate even if AI demand remains healthy.

Contrarian view: the consensus is treating Huang’s endorsement as validation, but endorsement is not a revenue stream. The more interesting signal is that the custom-chip cycle is still early, which means the stock with the most obvious long-duration optionality may actually be the one with the lowest execution bar and lower multiple — NVDA — not the one priced for perfection. What would falsify the cautious view is evidence over the next 1-2 earnings cycles that MRVL converts design wins into sustained margin expansion and accelerating backlog; absent that, the risk/reward remains skewed toward NVDA over MRVL.

Key Decisions for Investors

• Long NVDA / short MRVL as a 1-3 month relative-value pair: the trade benefits if AI capex remains strong but custom-silicon adoption ramps slower than the market is pricing. Target 10-15% relative upside; cut if MRVL starts showing durable revenue reacceleration or margin surprise.
• Use NVDA call spreads rather than outright MRVL longs to express AI upside: lower valuation leaves room for multiple support if growth merely normalizes, while MRVL needs both execution and sentiment to cooperate.
• If wanting exposure to custom silicon without single-name execution risk, watch MSFT and AMZN as the real demand anchors; any disclosure of accelerated in-house chip deployment is a warning sign for MRVL, but usually not an immediate negative for the broader AI hardware basket.
• Set an alert for MRVL post-earnings guidance and gross margin trajectory over the next 1-2 quarters; a miss on either is the clearest catalyst for multiple compression and a potential short entry on strength.