Nvidia Announces Full Production of Groq 3 LPX, Vera Rubin Boosts Inference Efficiency Several-Fold, SpaceXAI Doubles Down on Agents and Space AI

Nvidia stated that based on SemiAnalysis' AgentX workload and using the DeepSeek V4 Pro model, Vera Rubin achieves up to 30x the throughput per megawatt of the previous generation GB300 NVL72, with per-token costs reduced by up to 35x. SpaceXAI will adopt Vera CPUs to accelerate next-generation agentic AI applications and expand platform buildout based on Vera Rubin, with plans to extend an optimized Vera Rubin NVL72 into space for the first-generation Starmind AI satellites.