NVIDIA Vera Rubin NVL72
A next-generation AI acceleration platform from NVIDIA designed for data centers. Based on the Vera Rubin architecture, the NVL72 configuration integrates large-scale GPU nodes and supports rack-level liquid-cooled deployment.
Timeline (last 90 days)
vLLM provides day-0 support for NVIDIA Vera Rubin NVL72, reporting 7.8x throughput per GPU versus GB200 NVL72 (AgentX) and 3.7x VLM throughput versus GB300 (MLPerf v6.1), with locality domains tuned for MoE routing.
Giga Computing's exhibit includes a configured NVIDIA Vera Rubin NVL72 platform, Vera CPU and Groq 3 LPX nodes, and an OCP ORV3 rack covering compute, storage, management, and GPU systems.
ThunderKittens has been ported to the NVIDIA Vera Rubin NVL72 platform, with NVFP4 GEMM kernels reconstructed around the new hardware.
All Facts (3)
vLLM provides day-0 support for NVIDIA Vera Rubin NVL72, reporting 7.8x throughput per GPU versus GB200 NVL72 (AgentX) and 3.7x VLM throughput versus GB300 (MLPerf v6.1), with locality domains tuned for MoE routing.
50%UnverifiedGiga Computing's exhibit includes a configured NVIDIA Vera Rubin NVL72 platform, Vera CPU and Groq 3 LPX nodes, and an OCP ORV3 rack covering compute, storage, management, and GPU systems.
50%UnverifiedThunderKittens has been ported to the NVIDIA Vera Rubin NVL72 platform, with NVFP4 GEMM kernels reconstructed around the new hardware.
50%