Nvidia Unveils Nemotron 3 Ultra & Vera Rubin AI Platform

Nvidia has unveiled Nemotron 3 Ultra, an ultra-large open artificial intelligence model with 550 billion parameters. The company also announced it has started mass production of its next AI server platform, Vera Rubin. This move strengthens both AI models and data center infrastructure. Nemotron 3 Ultra is a large language model with 550 billion parameters. Nvidia plans to release it as an open model this week. Nvidia states it offers top-tier performance among open models developed by U.S. companies and demonstrates competitiveness against major open models, including some Chinese models. External assessments show Nemotron 3 Ultra scored 48 points on the Artificial Analysis intelligence index. This is above Google's Gemma 4 31B, but below the Chinese model Kimi K2.6. Nvidia emphasizes the model's practical usability, highlighting higher performance in tokens generated per second and better cost efficiency. The Vera Rubin platform is a data center AI server platform. It integrates high-performance storage and network systems with Nvidia's next AI GPU Rubin and its in-house AI CPU Vera. Nvidia states Vera Rubin provides higher processing efficiency for agent AI environments. This dual announcement indicates Nvidia's strategy to build an ecosystem from AI model development to data center infrastructure. This matters because it shows Nvidia's comprehensive approach to the evolving enterprise AI market.