NVIDIA reported revenue of $96.2 billion for the second quarter of fiscal 2027, ended July 26, 2026, up 18% from the previous quarter and up 106% from a year earlier, according to the company’s earnings release published August 26, 2026. GAAP and non-GAAP gross margins both came in at 75.0%, and GAAP diluted earnings per share reached $2.46, up 128% year over year, on net income of $59.7 billion.
The Data Center business remains the engine, posting $89.0 billion in quarterly revenue, up 18% sequentially and up 117% from a year ago. The company’s other reporting segment, Edge Computing, contributed $7.2 billion, up 13% from the previous quarter and up 27% year over year.
“AI has reached its inflection point. It’s doing useful work. Its tokens are productive and profitable. Now, compute is revenue,” said Jensen Huang, founder and CEO of NVIDIA.
Huang framed the demand side as broadening beyond the handful of buyers that drove the first wave of buildout, pointing to new AI labs and startups, multiple frontier labs scaling in parallel, an open-model ecosystem, and physical AI systems coming online. He tied the moment directly to the company’s newest platform.
“The AI infrastructure buildout is at full steam. Vera Rubin, now in full production, was built to power exactly this moment,” Huang said.
For the third quarter of fiscal 2027, NVIDIA guided revenue to $108.0 billion, plus or minus 2%, with GAAP and non-GAAP gross margins of 74.0%, plus or minus 50 basis points. The guidance carries a significant condition: the company is not assuming any Data Center compute revenue from China in its outlook. Operating expenses are expected to run approximately $9.2 billion on a GAAP basis and $9.0 billion non-GAAP.
The quarter also produced $21.3 billion in free cash flow, and NVIDIA returned approximately $26.0 billion to shareholders through buybacks and dividends, leaving roughly $99.0 billion remaining under its repurchase authorization. The balance sheet expanded to $320.3 billion in total assets, and the cash-flow statement shows the company took in $24.9 billion from a debt issuance during the quarter while deploying $42.4 billion into equity securities over the first half of the fiscal year.
NVIDIA’s Quarter of Racks, Power and Capital
The earnings cap a quarter in which the company moved aggressively on every input to the buildout: chips, financing, power and land. On August 10, 2026, NVIDIA announced strategic partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to establish independent compute financing platforms aimed at mobilizing over $500 billion of third-party capital for AI infrastructure, an effort Unite.AI covered as NVIDIA’s push to finance compute as an asset class. The partnerships rest on signed memorandums of understanding and remain subject to definitive agreements.
A week later, on August 17, 2026, the company secured land, power and shell capacity at the PORTS-Pike Technology Campus in Ohio through a partnership with SB Energy to host NVIDIA compute, and on August 21, 2026, it took a minority stake in data-center developer Cloverleaf Infrastructure. On the demand side, SpaceXAI committed to deploying NVIDIA’s Vera CPUs for its next generation of agentic AI applications, while Lancium signed on to deploy gigawatt-scale AI factories with the company. Rounding out the quarter, NVIDIA said Vera Rubin racks are now running at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius, and its edge business added the Jetson Orin Nano 2 robotics computer alongside an RTX Spark partnership with Microsoft for Windows PCs.
NVIDIA’s Second Quarter by the Numbers
- Total revenue: $96.2 billion, up 106% year over year
- Data Center revenue: $89.0 billion, up 117% year over year
- Edge Computing revenue: $7.2 billion, up 27% year over year
- GAAP gross margin: 75.0%, up from 72.4% a year ago
- GAAP diluted EPS: $2.46, up 128%; non-GAAP diluted EPS: $2.22, up 120%
- Operating income: $63.7 billion, up 124%
- Research and development spending: $7.1 billion, up from $4.3 billion a year ago
- Free cash flow: $21.3 billion
- Shareholder returns: approximately $26.0 billion in buybacks and dividends
- Third-quarter outlook: $108.0 billion, plus or minus 2%
One methodology note on the non-GAAP figures: beginning with the first quarter of fiscal 2027, NVIDIA’s non-GAAP measures no longer exclude stock-based compensation, and the historical comparisons in the release have been restated to match.
What Comes Next for NVIDIA
NVIDIA will pay its next quarterly cash dividend of $0.25 per share on October 1, 2026, to shareholders of record on September 10, 2026. The company was set to discuss the quarter on a conference call at 2 p.m. Pacific time on August 26, 2026, with a replay available through its third-quarter earnings call.
The forward line to watch is the China assumption. Excluding Data Center compute revenue from China entirely, NVIDIA still expects $108.0 billion in third-quarter revenue, which would represent 12% sequential growth and roughly 57% growth from the $68.9 billion implied year-ago quarter. The Vera Rubin ramp, now in full production, is the supply-side piece of that guide.
How Groq 3 LPX Extends Vera Rubin’s Inference Reach
The most architecturally significant product note in the release is that NVIDIA Groq 3 LPX, the company’s interactive AI inference accelerator, entered full production during the quarter. Groq 3 LPX is an extension of the Vera Rubin platform aimed at a specific bottleneck in agentic AI: agents generate massive volumes of tokens across hundreds or thousands of inference steps, so the rate of token generation for an individual user determines how quickly an agent completes each step of its work.
NVIDIA says Groq 3 LPX delivered a record 3,400 output tokens per second in Artificial Analysis benchmarking running the open-source Gemma 4 31B model with a 100,000-token context, and 4x faster responsiveness for latency-sensitive workloads than the nearest alternative platform. The rack configurations pair the accelerator with NVIDIA BlueField-4 DPUs, Vera CPU racks, Vera BlueField-4 STX storage and Spectrum-6 SPX Ethernet. The company also said Blackwell led across every category in the MLPerf Training 6.0 benchmarks and in AgentPerf, a new agentic AI infrastructure benchmark.
The branding carries its own disclosure: the Groq 3 LPX release notes that the Groq and LPU marks are used under license from Groq, Inc., and the earnings cash-flow statement shows a $2.9 billion payment to Groq, Inc. during the quarter. Nebius is the first AI cloud adopting Groq 3 LPX, bringing it to its Token Factory inference platform, with purpose-built inference cloud Groq planning to be among the earliest adopters to follow.
Credit: Source link
