Nvidia's Q2 FY2027 revenue hit $96.2 billion, up 106% year-over-year and beating the $92.2 billion consensus. CEO Jensen Huang declared "demand is accelerating," even as Q3 guidance of $91 billion fell short of the $103.9 billion analysts projected.
Key facts
- Q2 revenue: $96.2B, up 106% YoY
- Data center revenue: $89B vs. $85.7B est.
- Q3 guidance: $91B ±2%, below $103.9B consensus
- Non-GAAP EPS: $2.22 vs. $2.06-$2.09 est.
- Hyperscale: $48.7B; ACIE: $40.3B
Nvidia's fiscal second quarter results, reported Wednesday after market close, show a company still riding the AI infrastructure wave—but the forward guide reveals a market that may be starting to price in a slowdown that hasn't arrived yet. Revenue of $96.2 billion for the three months ended July 26 represented an 18% sequential gain and a 106% jump from the year-ago period, according to Fortune's report. Non-GAAP EPS came in at $2.22, beating the $2.06-$2.09 range analysts had modeled.
The data center segment—Nvidia's AI engine—delivered $89 billion in revenue, up from $75.2 billion last quarter and $39.1 billion a year earlier. The company's new segment breakdown splits the business into $48.7 billion in hyperscale revenue and $40.3 billion from AI clouds, industrial, and enterprise (ACIE) customers, a categorization designed to separate the mega-cloud buyers from AI-native clouds, sovereign AI, and on-prem deployments.
The tension sits in the guidance. Nvidia forecast Q3 revenue of $91 billion plus or minus 2%, which the company frames as beating "the average analyst expectation of $103.9 billion"—a curious framing, since $91 billion is nearly $13 billion below that consensus. The stock traded flat in after-hours action following a 1.6% intraday decline.
The guide is the story
Investors have grown accustomed to Nvidia sandbagging guidance and then blowing past it. The $91 billion midpoint implies a 5% sequential decline—an unusual signal from a company that has posted relentless growth. One read: hyperscaler purchasing is lumpy, and the company is being conservative ahead of the Vera Rubin platform ramp. Another read, less charitable: the demand curve Huang insists is "accelerating" may be flattening at the margin, and the company is managing expectations rather than surprising to the upside.
Notably, Nvidia began including stock-based compensation in its non-GAAP results this quarter, a change Fortune notes "makes direct comparisons to previous fiscal years less of an apples-to-apples distinction." The $2.22 non-GAAP EPS figure is not directly comparable to last quarter's $1.87 on the old basis.
The competitive backdrop
The results land a day after our own coverage of OpenAI's Jalapeño chip beating Nvidia's Blackwell on InferenceX, and days after Nvidia's $1 billion investment in Poolside at a $12 billion pre-money valuation, per our prior reporting. The company is simultaneously defending its silicon moat against custom silicon from OpenAI and Google while pouring capital into AI application layers. Nvidia also announced a 10-GW AI data center partnership with SB Energy in Ohio on August 22. The $96.2 billion quarter validates that strategy so far—the question is whether the $91 billion guide hints at a ceiling.
What to watch
Watch whether Nvidia beats its own $91 billion Q3 guide when it reports in November—and how the Vera Rubin ramp affects gross margins. Also track hyperscale vs. ACIE mix: a shift toward the latter would signal demand broadening beyond the largest cloud providers.
Source: fortune.com
