• NVIDIA says Vera CPU systems are now shipping at scale, with AWS receiving its first Vera server and Vera Rubin GPU
  • Earlier Vera systems have reached Oracle Cloud Infrastructure, Anthropic, OpenAI and SpaceXAI, moving the processor into customer qualification and integration

The Fact

NVIDIA says systems based on its Vera CPU are now shipping at scale, with Amazon Web Services receiving its first Vera server and Vera Rubin GPU. NVIDIA said Vera systems had previously been delivered to Oracle Cloud Infrastructure, Anthropic, OpenAI and SpaceXAI.

Vera is NVIDIA's first custom CPU and can operate in standalone servers or as the host processor in Vera Rubin NVL72 systems. NVIDIA designed it for CPU-heavy work around accelerated computing, including orchestration, data processing, tool calling and code execution. The company announced Vera in May with cloud providers, AI laboratories and server manufacturers preparing to adopt it. The latest deliveries move the processor beyond that launch stage, but NVIDIA has not said that the named customers are running production workloads on Vera or disclosed when AWS will make Vera-based infrastructure generally available.

The assessment

AWS now has Vera hardware in hand, which means it can begin the practical work of testing the system inside its own environment. That includes firmware, operating software, networking and the workloads it expects to run. Until those checks are complete, the servers are delivered hardware rather than usable cloud capacity.

AWS has said it plans to offer infrastructure based on Vera, but it has not given a date for general availability. NVIDIA has also named Oracle, Anthropic, OpenAI and SpaceXAI as recipients, without saying how many systems they received or whether any are already running production workloads. For BTW readers, the next important evidence will come from those deployments. Service-launch dates, qualification results and named workloads will show whether Vera is moving smoothly from delivered hardware into regular use.

What to watch

Watch for AWS to announce when Vera-based EC2 infrastructure becomes available and for Oracle, Anthropic, OpenAI or SpaceXAI to confirm that their systems are running production workloads. Details on system numbers and qualification timelines would show how quickly the first deliveries are moving into regular use. Vera Rubin deployments will also show how much demand is for standalone Vera servers versus the full rack-scale system.