News stories tagged with #Datacenter
NVIDIA Unveils Vera-CPU Rack at GTC 2026: New Benchmark for CPU-Only AI Infrastructure
At GTC 2026, NVIDIA unveiled the Vera-CPU Rack, a system designed for CPU-only inference featuring 256 Vera CPUs based on a custom Arm v9.2-A core called Olympus. Each CPU includes 88 cores with SMT, FP8 support, and up to 1.5 TB of RAM. The rack provides 400 TB of LPDDR memory and a total memory bandwidth of 300 TB/s, with individual chips achieving 1.2 TB/s. Optimized for agentic workloads, reinforcement learning, and AI training, the solution is being co-developed with partners like HPE, with the Cray GXC240 supporting up to 640 Vera CPUs per rack.
Nvidia Integrates Groq 3 LPU into Vera-Rubin Platform: A New Era of Low-Latency AI Inference Begins
At GTC 2026, Nvidia announced the integration of Groq’s 3rd-generation Language Processing Unit (LPU) into its new Vera-Rubin-NVL72 platform to dramatically boost AI inference throughput with ultra-low latency. Designed specifically for inference workloads, the LPU leverages high SRAM and internal bandwidth for rapid token processing. This technology complements Nvidia’s existing GPU ecosystem and is deployed in new LPX racks. Partners such as HPE and Giga Computing showcased next-generation AI factories and high-performance computing infrastructure at the event, built around these advancements.