Marvell Just Fell 10% After Record Earnings. Is This an AI Buying Opportunity?

By AltIndex Research · 9 min read · August 28, 11:33 am

Marvell just delivered record revenue, raised its forecasts for the next two years, and said AI-related bookings remain exceptionally strong. Investors responded by wiping roughly 10% off the stock.

Marvell Technology (MRVL) reported record quarterly revenue of $2.74 billion, up 37% year over year. Its Data Center business grew even faster, climbing 46% to $2.17 billion and accounting for nearly 80% of total revenue. Management also raised its fiscal 2027 revenue outlook to approximately $12 billion and now expects revenue to reach roughly $18 billion in fiscal 2028.

Yet MRVL shares fell 10.3% today. As with many AI stocks lately, the problem wasn't Marvell's earnings; it was sky-high expectations. After an extraordinary run in the stock and excitement surrounding Marvell's expanded custom-chip relationship with Google, investors were looking for an even bigger near-term payoff. Instead, management indicated that some of the larger contributions from those programs won't arrive until fiscal 2029.

That leaves investors with an interesting question: Has the selloff created a buying opportunity, or was Marvell simply priced too aggressively heading into earnings?

Marvell's earnings were strong almost everywhere you look

Strip away the stock reaction and Marvell's quarter looked strong. Revenue reached a record $2.739 billion, up 37% from a year ago and 13% sequentially. That also came in $39 million above the midpoint of Marvell's previous guidance.

The Data Center segment was the clear driver, reaching approximately $2.17 billion in revenue and growing 46% year over year. It now represents 79% of Marvell's total revenue, making the company's exposure to the AI infrastructure buildout increasingly difficult to miss. Profitability improved alongside revenue, with Marvell generating $308 million in GAAP net income, $865.9 million on a non-GAAP basis, and approximately $606 million in operating cash flow.

More importantly, management doesn't expect the growth to stop here. Marvell expects third-quarter revenue of approximately $3.15 billion, while its fiscal 2027 revenue forecast has been raised to roughly $12 billion. Management now expects approximately $18 billion in fiscal 2028 revenue as AI infrastructure demand continues to expand.

So why did Marvell fall 10%?

Marvell entered earnings with an unusually high bar. Even after Friday's selloff, the stock is still up roughly 181% over the past year and 165% over the past six months. That kind of performance changes what investors expect from an earnings report: beating estimates is no longer necessarily enough when the valuation already assumes years of exceptional growth.

One of the biggest sources of those expectations has been Marvell's expanded relationship with Google. The companies have broadened their custom-silicon work across several product categories, potentially giving Marvell a much larger role in Google's AI infrastructure. But investors looking for an immediate boost received a more nuanced answer on the earnings call.

Marvell said revenue from programs covered by the expanded agreement through fiscal 2028 is already incorporated into its outlook, while some of the larger programs now entering execution or production are expected to contribute much more significantly in fiscal 2029. The long-term opportunity may therefore have gotten larger, but some of the growth investors were hoping to add to their near-term models was already priced in.

The alternative data still looks surprisingly strong

That's where the selloff becomes more interesting. At AltIndex, we track dozens of traditional and alternative signals for Marvell, including employee business outlook, web traffic, audience growth and analyst expectations. Several of those signals suggest the underlying business momentum remains strong despite Friday's stock-market reaction.

Employee business outlook is one of the cleaner signals. In September 2024, roughly 76% of Marvell employees according to online reviews tracked by AltIndex had a positive business outlook. That figure reached 84% by mid-2025 and has since climbed to 92%. Improving internal confidence doesn't tell us exactly what Marvell will earn next quarter, but the direction is notable alongside accelerating Data Center revenue and management's increasingly ambitious forecasts.

Percentage of employees with a positive business outlook, based on employee-review data tracked by AltIndex.

Other signals are moving in the same direction. Our web-traffic estimates show approximately 364,000 monthly visits in July, up about 24% from July 2025, while Marvell's Instagram following has risen roughly 41% over the past year. Marvell also entered earnings with an AI Score of 77, comfortably above the Buy threshold of 60, with particularly strong audience and employment signals.

None of these metrics determines whether MRVL is cheap or expensive. They do, however, make it harder to argue that Friday's decline reflects a deterioration in the underlying business. The selloff appears much more closely tied to the timing and valuation of future AI growth.

Strong data doesn't automatically make Marvell cheap

This is the most important distinction for investors considering buying the dip. Marvell's fundamentals can continue improving while its stock falls if expectations were simply too high. Even after Friday's decline, Marvell still trades at a substantial premium to many semiconductor companies, with investors paying today for growth expected to arrive over several years.

That premium isn't necessarily irrational. Marvell expects revenue to grow around 45% this fiscal year, Data Center revenue is accelerating, its custom-silicon business is ramping, and management sees substantially more growth ahead. But Friday's reaction illustrates how little room for disappointment remains when a stock has already risen almost threefold in a year.

What would make the selloff more interesting?

For investors considering MRVL after the decline, the first thing to watch is whether Marvell's custom-chip pipeline begins showing up more visibly in revenue. Management expects its custom business to more than double next year, and delivering against that forecast would provide stronger evidence that design wins are translating into financial results.

Data Center growth is equally important. The segment already accounts for nearly four-fifths of Marvell's revenue and is expected to accelerate again next quarter. If that growth continues while employee business outlook, web activity and other alternative data signals remain strong, today's lower share price would become easier to justify.

October 6 could be the next big catalyst

Investors may not have to wait until the next earnings report for more clarity. Marvell is holding its Investor Day on October 6, when CEO Matt Murphy and the company's senior leadership team are expected to provide a deeper look at its long-term strategy and AI infrastructure opportunity.

After Friday's reaction, the custom-silicon roadmap will likely receive particular attention. More detail about the timing of Google's programs, other hyperscaler opportunities and Marvell's path toward $18 billion in fiscal 2028 revenue could materially change how investors value the stock. That makes October 6 an important date for anyone considering buying the post-earnings decline.

The bottom line

Marvell's 10% decline looks more like an expectations reset than a deterioration in its AI business. Revenue just hit a record $2.74 billion, Data Center sales grew 46%, management raised its longer-term forecasts, employee business outlook has climbed to 92%, and our AI Score remains high at 77.

But after a roughly 181% one-year run, Marvell still isn't obviously cheap. The selloff makes the stock more interesting, while the October 6 Investor Day could provide the next important piece of the puzzle: how quickly its expanding custom-chip pipeline actually turns into revenue. At roughly $217, investors are getting a better price — but they're still paying for a lot of AI growth that hasn't arrived yet.