NVIDIA Posts $96.2 Billion Quarter and Guides to $108 Billion Assuming No China Revenue

MakeCryippreferred on

NVIDIA reported $96.2 billion in revenue for its fiscal second quarter on Wednesday, up 106% from a year earlier and ahead of the roughly $92.3 billion analysts polled by Visible Alpha had forecast. Data Center sales, the business built around the chips that train and run AI models, brought in $89.0 billion of that total, up 117% from a year ago.

The stock, which had slipped during Wednesday’s regular session, reversed course after hours and traded higher once the results were out.

“AI has reached its inflection point. It’s doing useful work. Its tokens are productive and profitable. Now, compute is revenue. And demand is accelerating.”

NVIDIA CEO Jensen Huang said that in the company’s own announcement of the results.

Why a Zero Matters More Than the Beat

The more revealing number sat inside the guidance NVIDIA filed with the SEC. For the current quarter, NVIDIA told investors to expect $108 billion in revenue, also ahead of the roughly $104 billion analysts had penciled in, and said that figure assumes no Data Center compute revenue from China at all.

That is not because Washington still blocks the chips. The White House cleared H200 exports to China in December 2025, with formal licenses following in February, under a deal that requires NVIDIA to hand over 25% of that China revenue to the U.S. government.

China’s own buyers are the ones declining the hardware, steering purchases toward domestic chipmakers instead. NVIDIA is now building its guidance as if the world’s second-largest economy simply is not a customer, and it still expects growth to accelerate from here.

The company’s broader AI push also extends beyond hardware, with its Nvidia AI model efforts highlighting the growing importance of software and AI inference alongside its GPU business.

Who Else Is Filling the Order Book

Data Center’s 93% share of revenue understates how many different kinds of buyers sit behind that number. Alongside the hyperscalers, a cluster of former bitcoin miners has spent this year converting mining capacity into leased AI computing power, signing multibillion-dollar hosting contracts with cloud and AI customers. The shift from crypto mining to AI also comes as NVIDIA faces a Nvidia class action lawsuit over allegations related to its disclosures about crypto mining revenue.

It is a small slice of NVIDIA’s order book next to the hyperscalers, but a real and growing one, and a sign of how far the GPU shortage has pulled capital out of mining and into AI hosting.

NVIDIA’s own numbers hint at how tight that supply still is. “Customer forecasts point to NVIDIA’s growth doubling next year,” Chief Financial Officer Colette Kress said, even as the company itself is guiding to roughly 70% revenue growth for fiscal 2028, citing supply constraints rather than a shortage of orders.

The Next Print Will Test the Split

The real test comes when NVIDIA reports fiscal third-quarter results. If $108 billion holds without a dollar of China revenue, and Beijing keeps steering buyers to its own chipmakers instead of the parts Washington just cleared, that will confirm the split between the two AI-chip markets is structural now, not a temporary casualty of the trade fight.