Nvidia's fiscal second quarter produced the kind of headline numbers that used to settle an argument. Revenue reached $96.2 billion, more than double what it was a year earlier. Data Center revenue, the business that now defines the company, came in at $89 billion, up 117 percent from a year ago and up 18 percent from the prior quarter. Adjusted earnings per share hit $2.22, ahead of the roughly $2.09 analysts had expected. On paper, this was not a close call.
The stock did not treat it that way, at least not at first. Shares had already drifted lower into the print, closing the regular session down 1.6 percent as broader markets turned cautious ahead of the release. In the first half hour after the numbers came out, the stock hovered near flat to slightly lower, an odd response to a quarter this size. It took longer, as commentary from the call and the scale of the Data Center number circulated, for the shares to move decisively higher, implying a gain approaching 4 percent versus the regular close by the time this article was prepared.
That gap between the size of the beat and the market's hesitation to reward it immediately is the real story of this quarter. Nvidia proved, again, that AI infrastructure spending is showing up in its income statement. What it did not settle is whether the market believes that growth rate can continue at the same pace, especially once a specific and quantifiable piece of the business has been walled off from what the company is willing to guide to.
The Data Center Engine Is Still the Whole Story
Every meaningful driver of this quarter ran through Data Center. Gross margin held at 75 percent on both a GAAP and adjusted basis. Operating income rose 124 percent to $63.7 billion on a GAAP basis. Net income reached $59.7 billion, up 126 percent from a year ago, while the adjusted figure of $54 billion grew more slowly, up 118 percent, a reminder that GAAP and adjusted results are not moving in lockstep this quarter and should not be quoted interchangeably. The smaller Edge Computing business, which includes gaming, professional visualization, automotive and OEM sales, grew 27 percent to $7.2 billion, a healthy number that is nonetheless a rounding error next to Data Center's scale.
Nvidia returned $26 billion to shareholders through buybacks and dividends during the quarter and has $99 billion left on its existing authorization, alongside a next quarterly dividend of $0.25 per share payable in October. Those figures describe a company generating far more cash than it currently needs to fund its own growth, which is itself part of the bull case: capital intensity has not yet caught up with the scale of demand.
Jensen Huang, the company's founder and chief executive, framed the quarter in blunt terms in the earnings release: "AI has reached its inflection point. It's doing useful work. Its tokens are productive and profitable. Now, compute is revenue." That is a claim about durability, not just demand. It says the AI buildout has moved from a phase where customers were buying capacity on faith to one where that capacity is generating measurable returns. The results in this release support the demand half of that claim. They do not, on their own, prove the returns half, which is a question for Nvidia's customers as much as for Nvidia itself.
The China Number Nobody Can See
The most consequential line in the entire release is not a result. It is a guidance assumption. Nvidia's outlook for the current quarter, revenue of $108 billion plus or minus 2 percent, explicitly excludes any Data Center compute revenue from China. That is a deliberate, quantifiable exclusion tied to export restrictions, not a rounding assumption.
This cuts two ways. On one hand, it means the $108 billion guide is a conservative floor rather than a full accounting of what Nvidia could sell if China access were restored, which gives the bull case a specific, testable source of upside that has nothing to do with new product cycles or new customers. On the other hand, it means the headline "guide above expectations" framing understates how much of Nvidia's addressable market is currently closed off by policy rather than by competition or demand. Investors weighing the stock now have to price both the business Nvidia is guiding to and a separate, unresolved policy variable that could reopen or permanently shrink a meaningful revenue pool. That is a harder number to underwrite than a simple growth rate, and it is a reasonable explanation for why the market did not immediately reward a beat of this size.