Earn with Nexo
Nvidia’s Jensen Huang aims to dominate new CPU market with Vera chip
Nvidia projects nearly $20 billion in CPU revenue as its first custom processor ships to OpenAI, Anthropic, and SpaceX AI.
Nvidia’s new Vera CPU, unveiled at GTC in March 2026, represents Nvidia’s first purpose-built processor designed specifically for agentic AI workloads. During the company’s May 21 earnings call, CEO Jensen Huang and CFO Colette Kress projected nearly $20 billion in standalone CPU revenue for the current fiscal year, with a total addressable market opportunity of $200 billion tied to agentic AI orchestration.
What makes Vera different
The Vera CPU packs 88 custom Nvidia Olympus Arm cores and delivers memory bandwidth of 1.2 TB/s. Nvidia claims the chip delivers double the performance-per-watt compared to conventional processors.
For agentic workloads specifically, Vera is reportedly 50% faster than existing options. Enterprise data queries run up to 3x faster on the new architecture.
The chip integrates into Nvidia’s broader Vera Rubin platform alongside Rubin GPUs and BlueField-4 DPUs.
First deliveries already in the wild
In mid-May 2026, Ian Buck personally delivered the initial Vera CPU production systems to OpenAI, Anthropic, SpaceX AI, and Oracle Cloud.
Dell is also already shipping Vera CPUs in its PowerEdge server line as of mid-May 2026. Full production and broader customer delivery of the complete Vera Rubin superchip platform is expected in the second half of 2026.