NVIDIA reported income of $96.2 billion for the second quarter of fiscal 2027, ended July 26, 2026, up 18% from the earlier quarter and up 106% from a 12 months earlier, in accordance with the corporate’s earnings launch revealed August 26, 2026. GAAP and non-GAAP gross margins each got here in at 75.0%, and GAAP diluted earnings per share reached $2.46, up 128% 12 months over 12 months, on web earnings of $59.7 billion.
The Information Heart enterprise stays the engine, posting $89.0 billion in quarterly income, up 18% sequentially and up 117% from a 12 months in the past. The corporate’s different reporting section, Edge Computing, contributed $7.2 billion, up 13% from the earlier quarter and up 27% 12 months over 12 months.
“AI has reached its inflection level. It’s doing helpful work. Its tokens are productive and worthwhile. Now, compute is income,” mentioned Jensen Huang, founder and CEO of NVIDIA.
Huang framed the demand facet as broadening past the handful of patrons that drove the primary wave of buildout, pointing to new AI labs and startups, a number of frontier labs scaling in parallel, an open-model ecosystem, and bodily AI methods coming on-line. He tied the second on to the corporate’s latest platform.
“The AI infrastructure buildout is at full steam. Vera Rubin, now in full manufacturing, was constructed to energy precisely this second,” Huang mentioned.
For the third quarter of fiscal 2027, NVIDIA guided income to $108.0 billion, plus or minus 2%, with GAAP and non-GAAP gross margins of 74.0%, plus or minus 50 foundation factors. The steering carries a major situation: the corporate shouldn’t be assuming any Information Heart compute income from China in its outlook. Working bills are anticipated to run roughly $9.2 billion on a GAAP foundation and $9.0 billion non-GAAP.
The quarter additionally produced $21.3 billion in free money stream, and NVIDIA returned roughly $26.0 billion to shareholders by buybacks and dividends, leaving roughly $99.0 billion remaining below its repurchase authorization. The steadiness sheet expanded to $320.3 billion in complete belongings, and the cash-flow assertion reveals the corporate took in $24.9 billion from a debt issuance throughout the quarter whereas deploying $42.4 billion into fairness securities over the primary half of the fiscal 12 months.
NVIDIA’s Quarter of Racks, Energy and Capital
The earnings cap 1 / 4 by which the corporate moved aggressively on each enter to the buildout: chips, financing, energy and land. On August 10, 2026, NVIDIA introduced strategic partnerships with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs and KKR to determine impartial compute financing platforms geared toward mobilizing over $500 billion of third-party capital for AI infrastructure, an effort Unite.AI coated as NVIDIA’s push to finance compute as an asset class. The partnerships relaxation on signed memorandums of understanding and stay topic to definitive agreements.
Per week later, on August 17, 2026, the corporate secured land, energy and shell capability on the PORTS-Pike Expertise Campus in Ohio by a partnership with SB Vitality to host NVIDIA compute, and on August 21, 2026, it took a minority stake in data-center developer Cloverleaf Infrastructure. On the demand facet, SpaceXAI dedicated to deploying NVIDIA’s Vera CPUs for its subsequent technology of agentic AI purposes, whereas Lancium signed on to deploy gigawatt-scale AI factories with the corporate. Rounding out the quarter, NVIDIA mentioned Vera Rubin racks at the moment are operating at CoreWeave, Google Cloud, Microsoft Azure, Oracle Cloud Infrastructure and Nebius, and its edge enterprise added the Jetson Orin Nano 2 robotics pc alongside an RTX Spark partnership with Microsoft for Home windows PCs.
NVIDIA’s Second Quarter by the Numbers
Complete income: $96.2 billion, up 106% 12 months over 12 months
Information Heart income: $89.0 billion, up 117% 12 months over 12 months
Edge Computing income: $7.2 billion, up 27% 12 months over 12 months
GAAP gross margin: 75.0%, up from 72.4% a 12 months in the past
GAAP diluted EPS: $2.46, up 128%; non-GAAP diluted EPS: $2.22, up 120%
Working earnings: $63.7 billion, up 124%
Analysis and improvement spending: $7.1 billion, up from $4.3 billion a 12 months in the past
Free money stream: $21.3 billion
Shareholder returns: roughly $26.0 billion in buybacks and dividends
Third-quarter outlook: $108.0 billion, plus or minus 2%
One methodology be aware on the non-GAAP figures: starting with the primary quarter of fiscal 2027, NVIDIA’s non-GAAP measures now not exclude stock-based compensation, and the historic comparisons within the launch have been restated to match.
What Comes Subsequent for NVIDIA
NVIDIA can pay its subsequent quarterly money dividend of $0.25 per share on October 1, 2026, to shareholders of report on September 10, 2026. The corporate was set to debate the quarter on a convention name at 2 p.m. Pacific time on August 26, 2026, with a replay obtainable by its third-quarter earnings name.
The ahead line to observe is the China assumption. Excluding Information Heart compute income from China completely, NVIDIA nonetheless expects $108.0 billion in third-quarter income, which might symbolize 12% sequential development and roughly 57% development from the $68.9 billion implied year-ago quarter. The Vera Rubin ramp, now in full manufacturing, is the supply-side piece of that information.
How Groq 3 LPX Extends Vera Rubin’s Inference Attain
Probably the most architecturally important product be aware within the launch is that NVIDIA Groq 3 LPX, the corporate’s interactive AI inference accelerator, entered full manufacturing throughout the quarter. Groq 3 LPX is an extension of the Vera Rubin platform geared toward a particular bottleneck in agentic AI: brokers generate huge volumes of tokens throughout a whole bunch or 1000’s of inference steps, so the speed of token technology for a person consumer determines how shortly an agent completes every step of its work.
NVIDIA says Groq 3 LPX delivered a report 3,400 output tokens per second in Synthetic Evaluation benchmarking operating the open-source Gemma 4 31B mannequin with a 100,000-token context, and 4x sooner responsiveness for latency-sensitive workloads than the closest various 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 corporate additionally mentioned Blackwell led throughout each class within the MLPerf Coaching 6.0 benchmarks and in AgentPerf, a brand new agentic AI infrastructure benchmark.
The branding carries its personal disclosure: the Groq 3 LPX launch notes that the Groq and LPU marks are used below license from Groq, Inc., and the earnings cash-flow assertion reveals a $2.9 billion fee to Groq, Inc. throughout the quarter. Nebius is the primary AI cloud adopting Groq 3 LPX, bringing it to its Token Manufacturing unit inference platform, with purpose-built inference cloud Groq planning to be among the many earliest adopters to observe.
