Nvidia's Vera Rubin enters full production

Nvidia’s next AI system, Vera Rubin, has moved from roadmap language into manufacturing and sampling. CNBC reported on February 25 that CEO Jensen Huang had said in January the system was in full production, and Nvidia finance chief Colette Kress said on the company’s latest earnings call that the first Vera Rubin samples had shipped to customers earlier that week. Nvidia on May 20 also reported fiscal first-quarter 2027 Data Center revenue of $75.2 billion, up 92% from a year earlier, underscoring how much of the company’s current business still rests on demand for AI infrastructure. ### So what exactly is Vera Rubin? Vera Rubin is Nvidia’s next rack-scale AI system, designed to succeed Grace Blackwell later in 2026. CNBC reported the system uses 72 Rubin GPUs and 36 Vera CPUs, with Nvidia describing Vera as its CPU component and Rubin as the GPU architecture inside the broader platform. CNBC reported on March 18, 2025, that Huang introduced Vera Rubin at GTC as part of Nvidia’s annual release cadence, with systems expected to ship in 2026. (cnbc.com) By February 25, 2026, CNBC said the platform was due to ship in the second half of the year. ### What changed this year? January 2026 is the key date because that is when Huang said Vera Rubin was in full production, according to CNBC’s February reporting. (cnbc.com) That matters because it moved the product from a conference-stage announcement into volume buildout ahead of commercial shipments. February 25 added the next milestone. (cnbc.com) Kress told analysts that Nvidia had shipped its “first Vera Rubin samples to customers earlier this week,” CNBC reported after the company’s fiscal fourth-quarter results. That meant early hardware was already in customer hands months before broad availability. ### Why are investors watching this so closely? (cnbc.com) Nvidia’s May 20 earnings showed the current business is still expanding at a pace that keeps attention fixed on the next platform. The company said total revenue for the quarter ended April 26, 2026, was $81.6 billion, up 85% from a year earlier, while Data Center revenue reached $75.2 billion, up 92%. (cnbc.com) CNBC reported that data center now accounts for more than 91% of Nvidia’s sales, and Kress said the company expects every model builder and cloud provider to eventually deploy Vera Rubin. At GTC in March, Huang said he saw $1 trillion in orders for Blackwell and Vera Rubin systems through 2027. ### What is Nvidia promising on performance and power? (investor.nvidia.com) CNBC reported that Vera Rubin will use about twice as much power as Blackwell but is expected to deliver 10 times more performance per watt. Nvidia said the system contains 1.3 million components sourced from more than 80 suppliers in at least 20 countries. Dion Harris, Nvidia’s AI infrastructure head, told CNBC the company had been giving suppliers “very detailed forecasts” and said, “We’re in good shape.” That comment was tied to questions about whether Nvidia could line up enough parts for a system of that scale. (cnbc.com) ### Who is expected to get it first? Amazon Web Services, Google Cloud and Microsoft are the named cloud companies most closely associated with first deployments in recent coverage. (cnbc.com) The original 2025 launch coverage identified Microsoft, Google and Amazon as the cloud companies whose spending would help determine whether Nvidia’s next generation kept the AI buildout going. TechTimes later reported first deployments were expected at AWS, Google Cloud and Microsoft in the second half of 2026. The next concrete milestone is shipment. CNBC reported Vera Rubin systems are expected to ship in the second half of 2026, and Huang said at GTC in March that demand for Blackwell and Vera Rubin combined could reach $1 trillion through 2027. (cnbc.com 1) (cnbc.com 2)