NVIDIA shows inference momentum

NVIDIA’s latest quarter gave investors two things they wanted at once: bigger-than-expected numbers and a roadmap update that kept the next hardware cycle intact. On May 20, the company reported fiscal first-quarter revenue of $81.6 billion, up 85% from a year earlier, with data center revenue of $75.2 billion, up 92%. (investor.nvidia.com) The earnings release also pointed to continued demand tied to inference, not just training. Coverage of the results said Nvidia’s next-generation Vera Rubin platform is on track for production shipments in the second half of 2026, extending the company’s argument that customers still need more compute for serving AI models at scale. (investor.nvidia.com) The market reaction was mixed immediately after the release, but the broader read-through was supportive. (investor.nvidia.com) Several reports said the results eased some investor concern that AI spending was outrunning real demand, even as supply-chain reporting pointed to mounting strain underneath Nvidia’s pace of product launches. ### Which number mattered most in the quarter? The $75.2 billion data center figure was the quarter’s central number. (investor.nvidia.com) NVIDIA said that business rose 92% from a year earlier, while total company revenue reached $81.6 billion for the quarter ended April 26, 2026. Jensen Huang, Nvidia’s chief executive, used the release to frame that demand as part of a broader buildout of “AI factories.” In the company statement, Huang said agentic AI was “doing productive work” and “scaling rapidly across companies and industries.” (theverge.com) ### Why are investors focused on inference rather than only training? Vera Rubin was the key clue. Reporting after the earnings said Nvidia told investors the platform was on track for production shipments in the second half of this year and would offer materially higher inference throughput than prior systems. (msn.com) That matters because inference is the recurring workload once models are deployed. Nvidia did not break out inference revenue in its earnings release, but the roadmap language and management commentary were read by market participants as evidence that customers are still planning around larger serving fleets, not just one-off training clusters. (investor.nvidia.com) ### What did the company actually say about the next hardware cycle? (investor.nvidia.com) The company’s public materials did not provide a full Rubin shipment schedule in the earnings release itself, but follow-up coverage of the call said production shipments are expected in H2 2026. Nvidia has been laying out a faster cadence across Blackwell, Vera CPU and Rubin products, with DigiTimes describing that cycle as a three-tier roadmap that compresses the time suppliers have to prepare components and packaging. (theverge.com) That compressed cadence is central to the story because Nvidia’s revenue growth is arriving alongside a more demanding manufacturing schedule. DigiTimes reported that suppliers are being pushed to accelerate development, raise spending and manage higher quality risks as AI demand climbs. ### Where is the strain showing up? (investor.nvidia.com) DigiTimes said Nvidia’s record revenue and margins are masking pressure across the supply chain. The report said compressed product cycles and surging AI demand are forcing suppliers to move faster, with quality and execution risks rising as a result. That leaves Nvidia in a familiar position: strong demand at the top of the stack, and a narrow hardware ecosystem underneath it. The company’s results showed that customers are still spending heavily, but the supply-chain reporting suggested that real-time inference economics remain tied to how quickly Nvidia and its partners can deliver the next systems at volume. (digitimes.com) ### What comes next on the calendar? June 4 is Nvidia’s shareholder record date for its increased quarterly cash dividend, and June 26 is the payment date, according to the earnings release. (digitimes.com) The larger operating milestone remains Vera Rubin production shipments in the second half of 2026, which investors will watch for in Nvidia’s next roadmap updates and quarterly disclosures. (investor.nvidia.com)