NVIDIA Vera CPU Ships, Targeting Agentic AI Workloads

NVIDIA's first custom CPU, the Vera CPU, built specifically for agentic AI, has begun shipping at scale, with systems delivered to key players across the AI ecosystem.

This marks a significant move by NVIDIA (NASDAQ:NVDA) into the CPU market for AI, a segment experiencing new demands as AI models shift from answering queries to actively performing tasks.

NVIDIA Vice President of Hyperscale and HPC, Ian Buck, personally delivered Vera CPU systems to Oracle Cloud Infrastructure, Anthropic, OpenAI, and SpaceXAI, underscoring the strategic importance of this launch.

The announcement also details a major expansion of NVIDIA's collaboration with Amazon Web Services (AWS), including plans for 2 million additional NVIDIA GPUs and the integration of Vera CPU-based infrastructure into AWS.

Agentic AI requires substantial CPU power for tasks like orchestration, tool-calling, and managing long-context retrieval, areas where traditional CPU designs are less optimized.

Vera is engineered with 88 custom NVIDIA-designed Olympus cores and offers 1.2TB/s of memory bandwidth, aiming to deliver up to 1.8x faster per-core performance on these specific agentic AI workloads.

This architecture is designed to accelerate the entire AI factory by completing concurrent, real-time tasks more rapidly, thereby improving user response times and overall efficiency.

Oracle Cloud Infrastructure plans to deploy hundreds of thousands of NVIDIA Vera CPUs, positioning itself as the first cloud provider to adopt Vera at hyperscale for enterprise-grade agentic AI infrastructure.

StartupHub.ai data shows NVIDIA with a strong StartupHub score of 82/100, reflecting its dominant position in AI hardware, while Vera itself scores a 2/100, indicating its early-stage market entry as a new product category.

The introduction of Vera positions NVIDIA to capture a new segment of the AI infrastructure market, particularly as agentic AI workloads intensify CPU demands across cloud providers and AI labs like OpenAI and Anthropic.

Vera is also integrated into NVIDIA's broader ecosystem, including the Vera Rubin GPU and BlueField-4 DPU, contributing to a unified memory architecture for enhanced AI compute utilization.

This strategic hardware push by NVIDIA underscores the evolving needs of advanced AI, moving beyond pure GPU acceleration to address the critical role of specialized CPUs in the agentic AI era.