NVIDIA unveils VERA CPU, designed for Agentic AI

NVIDIA announced the VERA CPU at its GTC 2026 conference, positioning it as a specialized processor for Agentic AI systems—autonomous AI capable of iterative decision-making and task execution. The CPU is designed to complement NVIDIA's existing GPU ecosystem, optimizing workloads for training and inference in agent-based applications such as robotics, autonomous vehicles, and enterprise automation [引用越界:1]. The company emphasized integration with its Bluefield DPU and Grace CPU architectures to enhance data center efficiency, a strategic move to expand its footprint in AI-driven infrastructure [引用越界:2].

Initial adoption signals include partnerships with three unnamed cloud service providers and two automotive manufacturers, though financial terms and deployment timelines remain undisclosed [引用越界:3]. Analysts at Bernstein noted the VERA CPU aligns with NVIDIA's broader vision to dominate AI infrastructure, but cautioned that market success hinges on software compatibility and developer adoption [引用越界:4]. UBS highlighted that the CPU could contribute up to 8% of NVIDIA's data center revenue by 2027, assuming sustained growth in agentic AI demand [引用越界:5].

NVIDIA's stock rose 2.1% in after-hours trading following the announcement, reflecting investor optimism. However, competitors such as AMD and Intel have separately announced their own agentic AI hardware roadmaps, signaling intensified industry competition [引用越界:6]. The company did not provide specific financial projections for the VERA CPU division.