For most other companies, the results would be a home run. But for Nvidia, whose quarterly financials have become a litmus test for the AI boom, Wall Street wasn’t convinced. Shares fell more than 3% in after-hours trading as the chipmaker came up short in most of Wall Street’s most optimistic forecasts. The stock was trading lower premarket, down about 1.3%.

The market reaction is somewhat paradoxical: Nvidia’s core business is still booming, with the company reporting a jump in sales of more than 50%. However, the company narrowly missed data center revenue estimates.

“The miss on data center revenue weighs on the name in spite of the broader beat. Though Nvidia is forecasting $54 billion in revenue next quarter, traders may see this as a bearish catalyst given some on the Street had estimates as high as $63 billion,” said senior vice president of product and strategy at Direxion, Ryan Lee.

Investors largely ignored a similar miss last quarter, but this time, likely prompted by recent nervousness that the AI sector may be in a financial bubble, the stock faltered after the earnings report.

“Being priced to perfection leaves little room for error, and traders were left wanting more this quarter,” said Direxion’s head of capital markets, Jake Behan. “When any company trades at such high multiples, anything short of exceptional starts to look like a problem. Nvidia’s revenue forecast wasn’t bad, but it lacked the lofty upside the market was looking for.”

An investor vibe shift

Nvidia is the AI boom’s darling. The company’s valuation has been propelled to new heights as tech companies pour millions into AI infrastructure, skyrocketing demand for Nvidia’s AI chips. In July, the chipmaker became the first publicly traded company to achieve a $4 trillion market value.

The consequence of this industry-leading success is that the company’s performance is now seen as a proxy for the broader AI market. Investors, already wary of recent bubble concerns, and an MIT study that found that most companies haven’t realized meaningful gains from AI pilots, are hyperconscious of any signs of a dip in demand.

Bubble concerns could be bad for both Nvidia’s valuation and its customer base of cloud giants and well-funded AI startups, but its latest earnings don’t paint a picture of AI spending slowing down. In fact, CEO Jensen Huang said that the company expects to see $3 trillion to $4 trillion in AI infrastructure spending by the end of the decade.

“If you were waiting for clear signs of a slowdown in AI, you didn’t exactly get it,” Behan said. “This quarter shows Nvidia is still firmly in the game, navigating geopolitical turbulence and regulatory challenges while maintaining its leadership in the AI space.”

Instead, the market reaction may point to a vibe shift in how investors are viewing the AI sector. Over the past few years, investors have largely looked past minor misses and elevated valuations, treating rapid AI spending as a given and betting that demand would continue accelerating across hyperscalers and AI startups alike. But now, even small revenue misses or geopolitical hurdles, such as Nvidia’s uncertainty around China sales, are drawing attention, suggesting that investors are no longer willing to give the sector the same benefit of the doubt.

“The market has been used to overshooting,” Melissa Otto, head of visible alpha research at S&P Global, told Fortune. “Even though it was an okay number, and fundamentally the company is in good shape, the expectations are high, and the valuation is not cheap. So as we look forward the question is, Do the assumptions still hold, and could there be some earnings revisions?”

She noted that Nvidia’s guidance was in line with expectations from a sales perspective, but a little bit less than expectations in terms of margins. On the call, Nvidia chief financial officer Colette Kress also indicated the company plans to ramp up its expenses, something Otto said could prompt margins to either flatline or compress in the second half of the year, stalling some of the earnings growth.

“Over the past two years, Nvidia has absolutely defied gravity,” she said. “So the question is, Where does that go next? We’re trying to understand if we’re going to see yet another $200 billion in earnings revisions now. Based on the opex [operating expenses] and what we saw from the revenues this quarter, it does, at least for the near term, seem to be muted.”

Geopolitical difficulties

China is another sore spot for Nvidia. The company has counted on China sales for an extra boost to its numbers in the past, but regulatory uncertainty has prevented it from including any China revenue in its second-quarter results.

For months, Nvidia has been stuck in regulatory limbo over its H20 chips, which are subject to new U.S. export controls. Earlier this month, Nvidia and AMD struck a deal with the Trump administration to grant licenses in exchange for a 15% revenue-sharing arrangement on China chip sales.

Prior to the call, analysts had predicted that Nvidia would not allude to China revenue in the earnings report. During the call, CFO Colette Kress said the company recorded no H20 sales to China in the quarter because the 15% duty hasn’t been codified into regulation despite some customers receiving licenses in recent weeks.  

The company estimates it could ship $2 billion to $5 billion in H20s next quarter if restrictions ease, but none of that revenue is baked into its forecast. This could be a problem for Nvidia as access to the Chinese market could be critical for the company, with CEO Jensen Huang describing China as a $50 billion market this year alone.

The company also faces rising competition from domestic Chinese chipmakers like Huawei and Cambricon, which are catching up technologically and benefiting from local government support. Nvidia has warned in its filings that it may be effectively blocked from China’s data center market if it cannot gain regulatory approval. At the same time, China’s AI ecosystem, particularly open-source model development, is thriving. Huang emphasized that Nvidia was continuing to advocate for U.S. approval to bring its Blackwell chips to China, calling it a “real possibility” and a critical part of maintaining U.S. leadership in AI technology.

“China’s situation is a reminder that no matter how strong a company is, macro forces still matter—regulation, trade tensions, and global politics are now part of the equation,” said Kate Leaman, chief market analyst at AvaTrade.