NVIDIA의 에이전트 AI 인프라용 특화 Vera CPU가 주요 클라우드 제공자 및 AI l 전역에 걸쳐 대규모 배포 중입니다

AWS has received its first NVIDIA Vera CPU server and Vera Rubin GPU, hand-delivered in Seattle by NVIDIA Vice President of Hyperscale and HPC Ian Buck. AWS is the latest stop in Vera's rapid deployment across the AI ecosystem. Previously, Buck delivered Vera CPU systems to Oracle Cloud Infrastructure and three leading AI labs: Anthropic, OpenAI and SpaceXAI.

Ian Buck explained the significance of the moment. "Agentic AI is creating a new CPU moment in the AI factory — as models move from answering to acting, Vera is purpose-built to keep that work moving at scale." The underlying shift reflects how agentic AI places new demands on infrastructure. As models increasingly perform complex tasks beyond simple question-answering, every agentic sandbox, tool call, orchestration layer and long-context retrieval operation represents CPU work that GPUs alone cannot handle.

Vera is engineered specifically for this reality. The processor packs 88 custom NVIDIA-designed Olympus cores, delivers 1.2 terabytes per second of memory bandwidth and provides up to 1.8 times faster per-core performance on agentic AI workloads compared to traditional designs. This architecture allows work to complete more quickly under constant load, improving the efficiency of the entire AI factory and helping users get results faster.

At Oracle Cloud Infrastructure, the handoff took place at the Oracle AI Customer Excellence Center where Karan Batta, who leads overall product management, and Gary Miller, chief customer and partner success officer, toured the unboxed Vera CPU system. OCI plans to deploy hundreds of thousands of NVIDIA Vera CPUs beginning in 2026. Batta described the rationale: "Vera's architecture is purpose-built for high-throughput reasoning workloads, delivering the efficiency, density and footprint OCI needs to power the next generation of enterprise AI." OCI is the first cloud provider to deploy Vera at hyperscale.

At SpaceXAI's offices in Palo Alto, NVIDIA's team walked Elon Musk through the system, with Musk asking detailed questions about cores, memory layout and cooling. SpaceXAI is evaluating Vera for reinforcement learning workloads and agent-based simulation pipelines that drive its training stack. At OpenAI's Mission Bay headquarters, Sachin Katti, head of compute infrastructure, received the handoff on an open-air balcony as Buck walked through Vera's features and opened the system to reveal its interior. At Anthropic's offices in San Francisco, James Bradbury, head of compute, heard details on how the new CPU differentiates the system. Bradbury said, "Scaling compute is an important accelerant for the growth of models. We're excited to see Vera emerge as a promising part of the ecosystem when solving for agentic workloads."

Vera operates as part of NVIDIA's broader codesign story, working alongside the Rubin GPU, BlueField-4 DPU, Spectrum-X and MGX rack architecture. Vera also serves as the host processor for Vera Rubin NVL72, where it pairs with a pair of Rubin GPUs via second-generation NVIDIA NVLink-C2C. In these systems, Vera and Rubin share a unified memory architecture that keeps accelerated compute highly utilized, with Vera's fast CPU cores and interconnect handling orchestration, control and data movement at 2 times the energy efficiency of traditional infrastructure.