Marvell Stock Has Fallen 50%. The Amazon Fear Behind the Drop May Be Old News

The 50% drop looks more like a sentiment reset than a broken business

Marvell now looks like a contest between fundamentals and fear.

The stock has fallen about 50% from its 52-week high of 329.80, to around 163.40. A drop that steep often pushes investors from questioning valuation to fearing a structural break.

Strong results still argue the core story is intact

Marvell still delivered record third-quarter revenue of $2.075 billion, up 37% year over year, with 51.6% GAAP gross margin and $2.20 GAAP diluted income per share. That does not remove all risk, but it does make it harder to argue the business suddenly broke.

That tension fits what happened over the summer. Marvell was caught in a broad selloff across chip stocks as AI spending anxiety spread, then faced additional pressure after Erste Group downgraded the stock to Hold. Even so, that critique was mainly about valuation, not a sudden loss of momentum.

The most persistent company-specific worry was that Marvell could lose Amazon's next Trainium design win. That concern has lost some of its force, not least because Marvell and Amazon remain bound by a signed five-year agreement covering custom AI chips and related silicon. If earnings continue to back the growth story, this pullback may look more like a sentiment reset than the start of a broken thesis.

Marvell's latest numbers make the customer-concentration bear case harder to sustain

One operating update has done most of the work in weakening that fear: the results no longer look fragile enough to support an obvious customer-shock story.

Revenue and data-center growth still look healthy

In the last reported quarter, Marvell delivered Q4 FY2026 revenue of $2.22 billion, slightly ahead of consensus, with data-center revenue of $1.7 billion. Non-GAAP EPS was $0.80, up 33% year over year. Those are not the results of a business clearly losing footing with a key hyperscaler.

That matters because investors often react more quickly to bad news than to confirming data. Once a stock has fallen hard, even a disputed rumor can carry more weight than stronger operating results.

Why the Amazon narrative has lost prominence

The Trainium worry was always about concentration and dependency. The latest update points the other way. Marvell highlighted accelerating bookings and expanding attach across AI infrastructure, and management said bookings continue to grow at a record pace while expecting year-over-year revenue growth to accelerate each quarter in fiscal 2027.

If a major customer relationship had truly deteriorated, those signals would be harder to read. That does not eliminate all customer-specific risk, but it does make the old Amazon bear narrative look less aligned with the latest operating evidence.

If demand is still expanding, the real risk is execution and financing

If the Amazon scare is fading, the next question is what could still punish the stock.

The clearest answer is not an obvious collapse in demand. The evidence still points to expansion, with management citing accelerating bookings and expanding attach across AI infrastructure and expecting year-over-year revenue growth to accelerate each quarter in fiscal 2027. A more credible bull-case threat is slimmer execution margin, not an immediate breakdown in the business.

Balance-sheet pressure is the more tangible fundamental risk

That risk shows up on the balance sheet. Marvell's long-term debt of $4.961 billion at the April quarter was up 66.63% year over year. That does not mean the company is in trouble, but it does suggest less room for error if growth wobbles while investors become less forgiving of capital intensity.

Sentiment risk may matter more than narrative risk now

This is where behavioral risk may outweigh story risk. Marvell stock has averaged a 31% peak-to-trough drop during market shocks, versus 16% for the S&P 500. Its deepest drawdown was a 66% plunge during the 2008-2009 crisis, and the stock has also been hit harder during periods of sovereign and geopolitical stress.

That is the real trap for a high-growth semiconductor name: once the sell-off starts, recency bias and herd behavior can dominate even while the underlying business is still executing.

Why the next earnings report matters

The next checkpoint is Marvell's next quarterly earnings in late August. Recent results from major technology companies have also pointed to continued heavy investment in data centers and AI infrastructure, which keeps the backdrop from looking obviously hostile.

So the split is fairly straightforward: bulls still have demand support, while bears have a stronger case on stock behavior and resilience under stress. If August confirms that demand, fear can unwind faster than many investors expect.

MRVL looks more like a visibility trade than a bargain hunt

This is a visibility trade, not a bargain hunt. With a median analyst target of $250 and several recent targets ranging from $300 to $385, the market is still assuming Marvell can turn AI demand into another strong quarter when it reports in late August. The opportunity exists because price action has been driven as much by fear as by fundamentals, especially after a broad selloff across chip stocks.

What bulls still need to see

The bullish case is straightforward: management has pointed to accelerating bookings and expanding attach across AI infrastructure, and the stock needs that visibility to carry into the next report. The most useful updates will be forward-looking, not just a restatement of past strength.

What would weaken the case

The cleanest invalidation is not a dramatic change in story. It is weaker execution while the company carries long-term debt of $4.961 billion. If the next update also shows demand broadening more slowly than expected, the market's caution may prove more justified than it looks today.