Nvidia’s data center business, the engine behind the AI trade, brought in a record $75.2 billion for the quarter, up 92% from a year earlier, ABC News reported today. The company also generated $48.6 billion in free cash flow, up from $26.1 billion a year ago, and returned roughly $20 billion to shareholders through buybacks and dividends, Investing.com reported today.

That cash return got another boost. Nvidia raised its quarterly dividend from one cent per share to 25 cents and authorized a new $80 billion share repurchase program, TECHi reported yesterday.

Chief executive Jensen Huang framed the moment in sweeping terms, calling the current AI buildout “the single largest infrastructure expansion in human history” and saying demand has “gone parabolic,” BBC reported today. That is the sort of language that usually sends investors home happy. This time, it barely moved the needle.

The stock reaction looks less like disappointment and more like exhaustion. Nvidia has beaten revenue estimates for six straight quarters, yet closed lower after four of its last five earnings reports, Investing.com reported yesterday.

Kyle Rodda, senior financial analyst at Capital.com, called the quarter “a garden variety beat,” ABC News reported today. His point was simple enough: the results were strong, but not shocking, and they had been well telegraphed by the very strong numbers from AI hyperscalers earlier in earnings season.

Scale is part of the problem. Nvidia now makes up about 8% of the S&P 500 and carries a market value of around $5.3 trillion, BBC reported today. At that size, even a blowout quarter has less room to surprise. Gravity becomes a little more stubborn when the company in question is already this heavy.

Ruth Foxe-Blader of Citrine Venture Partners blamed the muted reaction on a “law of large numbers” effect, saying that without belief in continued parabolic growth, even excellent numbers struggle to excite investors, BBC reported today. Victoria Scholar of Interactive Investor made the other half of the argument: shares had already rallied ahead of the print, so investors “bought the rumour, sold the fact,” BBC reported today.

So the market’s message was not that Nvidia missed. It was that a beat on schedule is no longer enough on its own.

The first thing hanging over the stock is China. Nvidia’s $91 billion second-quarter forecast assumes zero data center compute revenue from China, Investing.com reported today. In the same quarter a year ago, China data center revenue was $4.6 billion, Investing.com reported today.

There has also been no real movement on the H200 front. In January, the Trump administration began allowing Nvidia to sell the chip to Chinese customers under certain conditions, BBC reported today. But the company has still not received approval from Chinese authorities, and no chips have been delivered, ABC News reported today; Investing.com reported yesterday.

Huang told CNBC he had “largely conceded” the Chinese data center market to Huawei, and Nvidia’s China revenue has fallen to roughly 5% of total sales from above 20% before export controls tightened, BBC reported today; Investing.com reported yesterday. Forrester senior analyst Alvin Nguyen argued Nvidia can still flourish without China, saying global AI demand outside that market is enough to sustain growth, BBC reported today.

The other concern is competition. Analysts flagged worries that major cloud customers are developing their own AI chips, which could gradually chip away at Nvidia’s grip on data center compute, BBC reported today. That remains a risk investors are pricing, not a proven hit to revenue. The line between “threat” and “trend” still matters here.

Even with those worries, the broader spending cycle does not look broken. Nvidia says annual AI infrastructure spending could reach $3 trillion to $4 trillion by the end of the decade, BBC reported today. Vontobel Asset Management said the four largest U.S. cloud operators have raised their 2026 capital expenditure expectations this year, with supply-chain data pointing to another material increase in 2027, Vontobel Asset Management reported yesterday.

That same Vontobel commentary said planning horizons for AI infrastructure have stretched to about 24 months, longer than the usual semiconductor cycle, because the hardware is harder to build and the spending is being committed further in advance, Vontobel Asset Management reported yesterday. As of this week, the shift from training large models toward inference and agentic AI workloads is also keeping compute demand broad, since each user action now burns more processing power than a one-off training run, Vontobel Asset Management reported yesterday.

That is why the stock reaction looks so peculiar. The underlying demand story is intact enough to support the business, but not so untouched that every strong quarter resets the valuation. Investors want more than durability. They want a reason to believe growth can still run faster than the market has already pencilled in.

For Nvidia, a beat is now table stakes. What would move the shares is evidence that growth can stay parabolic, not merely strong, and that no single geography or customer decision can put a ceiling on the story.

A reopening of China at scale would help. So would a demand signal that clearly exceeds even Nvidia’s own forecast, or a fresh set of hyperscaler capital-spending numbers that push the market to rethink just how big the AI buildout can get. Near-term catalysts include whether China ultimately clears H200 imports, whether cloud capex guidance rises again, and whether Nvidia’s next-generation Rubin architecture ships on schedule.

For now, the message from the market is blunt. Nvidia delivered record numbers and still lost some shine, because the bar has become absurdly high. That is not a business problem. It is a valuation problem with a very expensive address.