NVIDIA Groq 3 LPX Now in Full Production With World-Class Speed for Agentic AI
Full production means revenue timing: the LPX accelerator extends Vera Rubin NVL72 into the latency-sensitive agentic inference market, where Nvidia claims 4x faster agent responsiveness than the nearest alternative platform. Nebius is the first AI cloud deploying it via its Token Factory, with inference cloud Groq lined up next.