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

Nvidia CEO Jensen Huang highlighted Marvell as a potential next trillion-dollar company, but the article argues Nvidia is the better buy based on fundamentals. Street expectations call for Marvell revenue growth of 41% this year and 45% next year versus Nvidia’s 81% this year and 41% next year, and valuation is cited as ~22x forward earnings for Nvidia versus ~70x for Marvell. Net: despite the bullish endorsement for Marvell, the piece remains skewed toward Nvidia as the higher-growth, relatively cheaper option.

Analysis

This is less a stock-picker note than a positioning signal: the market is being asked to pay almost 3x the multiple for a business with slower near-term growth and a much less proven earnings stream. That makes MRVL vulnerable to “story premium” compression if the next two quarters don’t show accelerating design-win conversion, while NVDA remains the cleaner way to own AI capex because its revenue base is already monetizing the installed demand wave.

The second-order effect is that custom silicon does not need to kill GPUs to matter: it can siphon specific inference workloads and still expand overall AI spend by lowering cost per token for hyperscalers. That is constructive for AMZN and MSFT on cloud economics, but only if they actually retain the capex savings rather than reinvesting them into more compute. In other words, the real trade is not NVDA versus MRVL in isolation; it is whether hyperscalers keep increasing AI budgets fast enough to absorb both platforms.

Contrarian view: the consensus may be overreading Jensen’s endorsement as an immediate fundamental catalyst for MRVL. A CEO quote can lift the stock for days, but the durable driver is whether a few concentrated customers ramp production without margin leakage or delays. If custom-chip revenue remains lumpy, MRVL’s valuation is doing most of the work; if NVDA’s growth merely normalizes but stays above 40%-plus, its multiple can hold better than skeptics expect.

Key Decisions for Investors

• Long NVDA / short MRVL as a relative-value pair for 1-3 months; thesis is that NVDA’s growth-to-multiple ratio remains superior and MRVL’s premium is fragile unless it prints clear custom-chip acceleration.
• Avoid chasing MRVL on endorsement-driven strength; wait for evidence of meaningful revenue contribution from custom silicon and gross margin expansion before considering a long.
• Use AMZN and MSFT as indirect AI-capex beneficiaries: long the hyperscalers against a basket of semis only if upcoming cloud commentary confirms that custom silicon lowers unit costs without slowing spend.
• If trading options, prefer NVDA call spreads over outright MRVL calls for a 3-6 month horizon; NVDA has the better risk/reward if AI demand stays intact but valuation resets remain a risk.
• Falsifier for the short-MRVL thesis: two consecutive quarters of upside revenue revisions and margin accretion from custom chips; until then, treat any multiple expansion as sentiment-driven and vulnerable to reversal.