Demand for AI infrastructure is accelerating rather than cooling, judging by Nvidia’s latest quarter. The company posted revenue of $96.2B for the three months through late July, up 18% sequentially and more than double the year-ago figure, with chief executive Jensen Huang describing an industry where compute itself has become the revenue generator.
Data center sales accounted for $89B of the total, a 117% gain from a year earlier. Huang credited a broadening customer base, noting that one lab drove last year’s buildout while today multiple frontier labs, startups and open-model developers are scaling at the same time.
The company guided third-quarter revenue to $108B, plus or minus 2%, with gross margin around 74%. Its forecast carries no data center compute sales from China, a reminder of the export-control constraints still hanging over the business.
Shareholder returns totaled $26B in the quarter, leaving about $99B in the buyback authorization. A dividend of $0.25 per share arrives October 1.
Vera Rubin production ramps are running at CoreWeave, Google Cloud, Microsoft Azure and Oracle Cloud Infrastructure, Nvidia said, as hyperscalers keep spending heavily on AI factories.
