All articles tagged with #vera cpu
Nvidia unveiled its Vera CPU for agents, aiming at agentic AI and a new $200 billion total addressable market beyond its GPUs, as hyperscalers and tech giants invest in advanced AI workloads, signaling further upside for the AI leader.
Nvidia used Computex to outline a new wave of AI products, updating its Vera CPU, Spark superchip and Rubin platform and signaling a shift toward agentic and physical AI designed to power AI agents, robots, autonomous vehicles and industrial systems, backed by partnerships with Cadence, Siemens, Synopsys, CrowdStrike and Palantir to broaden AI hardware and software adoption across data centers, PCs, cybersecurity and enterprise AI.
NVIDIA introduced Vera, its first standalone data-center CPU, claiming it runs AI workloads 1.8 times faster than x86 chips; production starts in Q3 as Nvidia expands beyond GPUs into the AI infrastructure stack. Vera will compete with Intel Xeon, AMD Epyc and AWS Graviton, and Nvidia also expanded its DSX software for energy efficiency; Dell will offer DGX Station on Windows later this year, with Anthropic, OpenAI and SpaceX named among Vera’s first major users.
NVIDIA unveiled Vera, a purpose-built CPU for AI agents that NVIDIA claims delivers up to 1.8x faster agentic workloads than x86, with 88 Olympus cores, 1.2TB/s memory bandwidth and NVLink-C2C for CPU-GPU cohesion; designed for standalone Vera servers, Vera Rubin systems and Vera BlueField-4 STX platforms, backed by a broad ecosystem of system builders and cloud providers, with early interest from NYSE, Anthropic, OpenAI, SpaceXAI, ByteDance and OCI; shipping this fall as part of a shift toward tokens-per-dollar economics in AI factories.
NVIDIA unveils the Vera CPU, the first processor purpose-built for agentic AI and reinforcement learning, claiming about twice the efficiency and 50% faster performance than traditional CPUs. Designed for rack-scale AI factories, Vera supports large-scale concurrent workloads (e.g., 256-core racks), uses high-bandwidth NVLink-C2C interconnects, LPDDR5X memory up to 1.2 TB/s, and features 88 Olympus cores. Backed by broad ecosystem partners and major customers, Vera is set for full production in the second half of 2026.
NVIDIA has unveiled its next-gen Rubin GPU architecture, along with Rubin and Blackwell GPUs and the latest Vera CPU. The Rubin GPUs, expected to be mass-produced by late 2025, will feature HBM4 memory and a 4x reticle design. The Blackwell GPUs will launch in 2024, followed by an Ultra version in 2025. The Vera CPU, based on TSMC's 3nm process, will be part of the GR200 Superchip module. NVIDIA aims to enhance power efficiency and AI capabilities with these new products.