Nvidia's Q2 Revenue Hits $96.2 Billion, Tops Estimates With $108 Billion Q3 Guidance
News Summary
Nvidia delivered another record-breaking quarter, reporting second-quarter fiscal 2027 revenue of $96.2 billion, up 106% from a year earlier and comfortably ahead of Wall Street's $92.3 billion consensus estimate. The chipmaker also issued a third-quarter revenue outlook of $108 billion, plus or minus 2%, signaling that demand for its AI computing hardware continues to accelerate rather than plateau. The results, announced after market close on Tuesday, August 25, 2026 (Eastern Time), pushed Nvidia shares higher in after-hours trading before the stock settled to a gain of more than 4% the following session.
Data Center Segment Leads the Charge
Nvidia's Data Center business, which houses its AI training and inference chips, generated $89.02 billion in the quarter, a 117% year-over-year increase. Company executives attributed the surge to accelerating deployments of the Blackwell Ultra architecture, with major cloud and hyperscale computing customers more than doubling their purchases compared with the same period last year. Networking products, which connect thousands of GPUs into large computing clusters, also contributed meaningfully to the segment's growth as data center operators build increasingly large AI training systems.
Earnings Per Share and Profitability
Adjusted earnings per share came in at $2.22, up 120% from the prior year and above analyst expectations of $2.09. Both GAAP and non-GAAP gross margins for the quarter stood at 75.0%. Looking ahead, Nvidia guided third-quarter gross margin to approximately 74%, with fourth-quarter margin expected to settle between 71% and 72%. The company said the moderation reflects higher memory component costs across the industry, which have risen faster than previously anticipated as demand for high-bandwidth memory used in AI accelerators continues to outpace supply.
Leadership Commentary
"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 chief executive officer of Nvidia, during the earnings call. Huang and other executives emphasized that enterprise and sovereign AI customers are increasingly deploying Nvidia's platforms for real-world applications, from scientific research to industrial automation, rather than purely experimental workloads. Management pointed to sustained order backlogs extending into next year as evidence that the current growth cycle has staying power beyond a single product generation.
Looking Toward Future Product Generations
Executives also outlined the revenue potential of upcoming architectures. The company estimated a roughly $25 billion revenue opportunity per major customer deployment cycle with the current Blackwell generation, rising to an estimated $40 billion opportunity with the next-generation Vera Rubin platform, once CPUs, GPUs, networking equipment, and software are included. This framing underscores Nvidia's strategy of selling complete AI computing systems rather than individual chips, a shift that has broadened its addressable market well beyond graphics processors.
Market and Industry Reaction
Investors and analysts covering the semiconductor sector said the results reinforce the view that large-scale investment in AI infrastructure remains in an early, expansionary phase rather than a mature or slowing one. Industry observers noted that the strong data center growth also reflects continued capital spending commitments from major cloud providers, who have signaled multi-year plans to expand AI computing capacity. The guidance for the current quarter, if achieved, would mark another sequential revenue record for the company.