Marvell posted record Q1 revenue, beat on earnings, and called demand for its custom AI chips exceptional. The stock fell.
Not for the first time. This is now the playbook: good results, sharp guidance deceleration, and the stock sells off anyway. In Q2 of fiscal 2026, the stock dropped 18% on a similar setup. The investor day got pushed back in May - 10% off. May 15th, the stock fell 6% on sector profit-taking. April 27th, another 5.3% on a rumor. The pattern is the story. At this multiple, there is no bad-news discount to find. Everything is already bought, and the market is just waiting for growth to confirm it.
The real variable isn't the beat. It's the deceleration.
Marvell reported $2.418 billion in Q1 FY27 revenue, up 28% year-over-year, with non-GAAP EPS of $0.80 - both beating expectations. That's the headline. But Q2 guidance was $2.20 billion, implying just 5.8% year-over-year growth. That's a cliff from 28% to under 6% in one quarter.
Why does Q2 slow? Seasonality for the industry, yes - Q2 is historically softer. But the deceleration ratio matters more than the excuse. A company trading at 54 times forward earnings can't have a 5x slowdown in growth and expect the market to shrug. The price already assumes acceleration, not a step-off.
The multiple is doing the work. The earnings can't catch up.
Forward EPS for fiscal 2027 sits around $3.83. At $208, that's 54 times next year's earnings. For context, the stock has already gained over 134% year-to-date. The math only works if FY28 and beyond deliver compounding that justifies front-loading two years of growth into today's price.
GARP doesn't apply when the multiple is 54x. GARP means growth at a price where the downside is bounded and the math is on your side - 13x on a 30% grower, 20x on a 40% grower. This is premium growth pricing. The thesis shifts from "is the growth real" to "is the growth fast enough, fast enough, fast enough to justify what you've already paid."
The Celestial AI deal adds optionality, not near-term certainty.
Marvell completed its $3.25 billion acquisition of Celestial AI in February, targeting optical interconnects for next-generation data center scale-up architectures. That's a real strategic move - optical connectivity is the next bottleneck as AI clusters grow beyond single-rack designs. But acquisitions of this size take time to integrate and rarely contribute to the next quarter's revenue. The market can't price in a technology shift that hasn't shipped yet, and shouldn't try.
What would break the setup.
The bull case holds if Q2 is a pure seasonality blip and H2 reaccelerates to 25%+ growth. If data center revenue keeps expanding at the 40%+ rates management hinted at in full-year FY27 guidance, the stock earns its premium. But if Q2 turns into Q3, and 5.8% becomes a trend rather than a trough, the 54x multiple becomes a 54x anchor.
The key risk isn't AI demand. It's the expectation gap. At this valuation, the market is pricing in flawless execution on custom silicon design wins, continued hyperscaler capex growth, and seamless integration of Celestial AI. Miss any one, and the re-rating goes the wrong way.
At 54x forward earnings, Marvell is no longer a GARP play. It's a conviction play. The growth is real. The AI infrastructure tailwind is real. But the price already reflects a future that hasn't arrived - and the Q2 guidance says that future might take longer than the stock allows.