Nvidia is sharpening the market story around Vera, its data-center CPU for agentic AI workloads, after reporting another record quarter.

The company said fiscal first-quarter data-center revenue reached $75.2 billion, up 92% from a year earlier. In the same release, Nvidia highlighted Vera as part of the Vera Rubin platform and described it as a processor purpose-built for agentic AI, alongside new networking and storage infrastructure for AI factories.

The more notable claim came on the earnings call. Nvidia framed Vera not merely as a companion to its GPUs, but as an entry into a separate CPU market tied to AI agents and reinforcement-learning workloads. The earnings transcript says Nvidia sees a $200 billion total addressable market for CPUs as customers move toward agentic systems.

That framing matters because it shifts part of the AI hardware debate away from accelerator supply alone. Agentic systems can place heavier demands on orchestration, memory movement, tool execution, and sequential control logic, all areas where CPUs remain central even inside GPU-heavy data centers.

The conservative read is that the $200 billion figure is Nvidia's market estimate, not booked revenue. Vera is also already part of a broader platform story that Nvidia introduced at GTC, so this is not a new product launch. What changed this week is that Nvidia put a concrete market size around the CPU opportunity and tied it directly to its earnings narrative.