NVIDIA Vera CPU: The Ultimate Processor for Agentic AI and Reinforcement Learning
NVIDIA recently introduced the NVIDIA Vera CPU, a revolutionary processor designed specifically for the era of agentic AI and reinforcement learning. This cutting-edge CPU is capable of delivering results that are twice as efficient and 50% faster than traditional rack-scale CPUs, making it a game-changer in the world of data processing, AI training, and agentic inference.
As the field of reasoning and agentic AI continues to evolve, the demand for high-performance, energy-efficient processors has never been greater. The NVIDIA Vera CPU builds on the success of the NVIDIA Grace CPU and is set to enable organizations of all sizes and across various industries to build AI factories that can handle agentic AI at scale. With its superior single-thread performance and bandwidth per core, the Vera CPU is in a class of its own, offering higher AI throughput, responsiveness, and efficiency for large-scale AI services.
Leading hyperscalers such as Alibaba, CoreWeave, Meta, and Oracle Cloud Infrastructure are already collaborating with NVIDIA to deploy the Vera CPU. Additionally, major system manufacturers like Dell Technologies, HPE, Lenovo, and Supermicro have embraced the Vera CPU, making it the new industry standard for critical AI workloads. This widespread adoption of the Vera CPU is set to democratize access to AI and drive innovation across various sectors.
The NVIDIA Vera CPU is configurable for every data center, with a new Vera CPU rack capable of integrating 256 liquid-cooled Vera CPUs to support over 22,500 concurrent CPU environments. This allows AI factories to deploy and scale to tens of thousands of simultaneous instances and agentic tools within a single rack. The Vera CPU rack is built using the NVIDIA MGX modular reference architecture and is supported by a global network of 80 ecosystem partners.
As part of the NVIDIA Vera Rubin NVL72 platform, Vera CPUs are paired with NVIDIA GPUs through NVIDIA NVLink-C2C interconnect technology, providing high-speed data sharing between CPUs and GPUs with 1.8 TB/s of coherent bandwidth. This platform also includes new reference designs that use Vera as the host CPU for NVIDIA HGX Rubin NVL8 systems, facilitating data movement and system control for GPU-accelerated workloads.
Vera systems partners offer both dual and single-socket CPU server configurations that are ideal for a range of workloads, including reinforcement learning, agentic inference, data processing, storage management, and high-performance computing. These systems integrate NVIDIA ConnectX® SuperNIC cards and NVIDIA BlueField®-4 DPUs for accelerated networking, storage, and security, essential for agentic AI applications.
The NVIDIA Vera CPU is designed for agentic scaling, combining high-performance, energy-efficient CPU cores, a high-bandwidth memory subsystem, and the second-generation NVIDIA Scalable Coherency Fabric. With 88 custom NVIDIA-designed Olympus cores, Vera delivers exceptional performance for a wide range of AI tasks, making it ideal for multi-tenant AI factories running numerous jobs simultaneously.
Leading companies like Cursor and Redpanda are already leveraging the power of the NVIDIA Vera CPU to enhance performance and efficiency in their AI applications. National laboratories and cloud service providers are also planning to deploy Vera CPUs, recognizing the transformative impact these processors can have on their operations.
The NVIDIA Vera CPU is now in full production and will be available from partners in the second half of this year. This groundbreaking processor is set to revolutionize the world of AI and reinforcement learning, driving innovation and scalability across industries. To learn more about the NVIDIA Vera CPU and its impact on the future of AI, watch the GTC keynote from NVIDIA’s founder and CEO, Jensen Huang, and explore the sessions available.
For more Information, Refer to this article.
