Marvell stock sinks 8%: is its $120B Google deal taking too long to pay off?

AI Sentiment: 28/100 Bearish
This score is generated through AI-driven analysis of the article's content.
powered by
Buy MRVL. The quarter beat and data-center revenue jumped 46% YoY, with management guiding data-center growth >60% in FY2028. The selloff is mostly timing: investors wanted Google’s incremental custom-chip revenue to show up immediately, but management says much of FY2028 is already embedded and the bigger incremental impact ramps in FY2029+. That’s a valuation reset, not an AI demand reset. Key risk: Google program ramps slower than expected (or gets delayed), so FY2029+ incremental revenue never materializes and the market’s “already-paid-for” premium stays unjustified.
Key Risk: Google custom-chip revenue ramps later/slower than expected, so FY2029+ incremental growth disappoints and the premium valuation can’t be earned.
Sell MRVL relative to NVIDIA/AMD (go long NVDA or AMD, short MRVL). The market is punishing MRVL for forecast-model uncertainty around the Google deal, while NVDA/AMD are less dependent on one mega-customer’s ramp timing for near-term narrative. If the AI capex cycle stays intact, MRVL’s fundamentals (data-center growth) should re-rate once investors get October 6 investor-day clarity. Key risk: MRVL’s custom AI accelerator share gains stall versus peers, so the relative gap widens and MRVL underperforms even after guidance detail.
Key Risk: MRVL loses share or margins in custom AI accelerators, so even with better disclosure it can’t catch up to NVDA/AMD performance.
- Marvell falls nearly 8% as Google revenue timing disappoints Wall Street.
- Fiscal 2029 emerges as the key year for Marvell's huge Google opportunity.
- Marvell's AI growth remains intact despite the sharp post-earnings selloff.
Marvell Technology stock NASDAQ:MRVL sank nearly 8% in Friday premarket trading as investors looked past another solid quarter and focused instead on when its enormous Google AI-chip opportunity will materially lift revenue.
Fiscal second-quarter revenue rose 37% to $2.74 billion, beating expectations, while Marvell raised fiscal 2027 revenue guidance to about $12 billion and fiscal 2028 to roughly $18 billion.
Those upgrades were not enough for a stock that had already gained about 184% in 2026. The problem was timing: investors wanted the Google agreement to force much larger near-term estimate increases.
Wall Street expected Google to show up sooner
The most revealing exchange came on Marvell’s earnings call.
JPMorgan analyst Harlan Sur asked why the potentially $120 billion Google program was not showing up more clearly in fiscal 2028 expectations.
“I actually thought that that would start to show up in FY 2028,” Sur said, according to the call transcript carried by MarketBeat.
He noted that Marvell still appeared to be guiding towards only about $5 billion to $6 billion of custom-chip revenue in calendar 2027.
CEO Matt Murphy responded that revenue tied to programs already underway is included in next year’s numbers, but said newer programs should contribute much more significantly in fiscal 2029.
That distinction helps explain Friday’s selloff.
Investors had treated the Google agreement as a reason to raise Marvell’s growth trajectory immediately. Management instead suggested that much of the existing fiscal 2028 benefit was already embedded in forecasts.
The market had effectively pulled part of Google’s future success into Marvell’s current valuation before the revenue arrived.
Fiscal 2029 is becoming the critical year
Murphy told analysts that the larger incremental impact should emerge in fiscal 2029 and beyond, while promising more detail at Marvell’s October 6 investor day.
That left analysts trying to bridge an unusually large modelling gap.
Melius Research analyst Ben Reitzes said investors were “just wrestling with this” and added that “these numbers are huge.”
His point was straightforward. A potential $120 billion opportunity spread over roughly six and a half years implies an enormous annual revenue pool, but investors still do not know how much is incremental, when individual programs ramp or what margins they will carry.
Marvell has therefore given the market the size of the prize without yet providing enough detail about the path to collecting it.
An 8% selloff does not break the AI thesis
The disappointing reaction does not mean Marvell’s underlying AI business has weakened.
Data-center revenue rose 46% year over year to $2.17 billion in the quarter, while management expects data-center sales to grow more than 60% in fiscal 2028.
TECHnalysis Research chief analyst Bob O’Donnell told Reuters that expectations around custom AI accelerators are “getting ahead of themselves.”
That explains how a beat-and-raise quarter can still produce a sharp decline.
Citi said it was encouraged by Marvell’s data-center outlook, while Morgan Stanley noted that the stock tends to react more to changes in long-range forecasts than individual quarterly beats.
Melius also raised its price target to $350 from $325 on Friday, arguing that Google, Microsoft and AI connectivity remain significant long-term opportunities.
The Google deal may ultimately prove as large as bulls expect. The near-term problem is that investors have already paid for part of that future.

Nvidia, AMD, Intel stocks slide premarket: why investors are suddenly taking profits

Micron stock has surged 220%: here’s the risk bulls may be underestimating

Nikkei 225 catches Nvidia’s tailwind, but KOSPI finds a reason to sell instead

Samsung stock retreat again: why a massive payout isn’t enough for investors

Why Marvell stock is slipping despite a classic ‘beat and raise’
No results found
Loading articles...
Failed to load articles. Please try again.