NVIDIA Groq 3 LPX Full Production

At Hot Chips 2026, Nvidia announced that its Groq 3 LPX AI inference accelerator is now in full production. The chip extends the Nvidia Vera Rubin platform, dramatically increasing token generation rates for agentic AI workloads. In Artificial Analysis benchmarking, Groq 3 LPX achieved a record 3,400 output tokens per second running Gemma 4 31B with a 100,000-token context, the fastest performance ever recorded for that model. Nebius Group is the first AI cloud to adopt the chip, planning to integrate it into its Nebius Group Token Factory production inference platform. Groq, the AI inference cloud, also plans to be among the earliest adopters. The announcement underscores Nvidia's execution on its AI roadmap, following mass production of Vera CPUs and Vera Rubin servers.