Nvidia Begins Vera CPU Deliveries to AI Clients as Oracle Plans 2026 Deployment
Nvidia has started shipping its Vera CPU systems to Anthropic, OpenAI, Oracle Cloud Infrastructure and SpaceXAI, with Oracle planning to deploy hundreds of thousands of units in 2026. Only 35% of investors now name Nvidia their top AI pick, down from 70% last July, while its 19% YTD gain trails peers.
1. Vera CPU Shipments to AI Clients
Nvidia has commenced delivery of its next-generation Vera CPU systems to early customers including Anthropic, OpenAI, Oracle Cloud Infrastructure and SpaceXAI. Executive Ian Buck personally handed over the first units at Anthropic’s San Francisco offices and at SpaceXAI’s Palo Alto facility.
2. Integrated Computing Platform
The Vera CPU forms part of a co-designed stack alongside Rubin GPUs, BlueField 4 DPUs, Spectrum-X networking and MGX rack architecture. In the Vera Rubin NVL72 configuration, Vera acts as host processor connecting to two Rubin GPUs via second-generation NVLink-C2C.
3. Oracle’s Hyperscale Deployment Plans
Oracle Cloud Infrastructure will be the first hyperscale cloud provider to roll out Vera CPUs, with plans to deploy hundreds of thousands of units beginning in 2026. This large-scale rollout underscores Nvidia’s strategy to embed CPUs across major cloud and AI infrastructure.
4. Investor Sentiment and Earnings Outlook
Investor conviction in Nvidia as a top AI logic pick has fallen to 35% from 70% a year ago, and its 19% year-to-date return trails competitors like AMD and Micron. Despite several analysts raising price targets ahead of the upcoming earnings report, shares have dipped on four of the last five quarterly results even after beats.