Nvidia Launches Groundbreaking Groq 3 LPX Chip for Real-Time AI Processing

Nvidia has officially started shipping its Groq 3 LPX inference accelerator, marking a significant advancement in AI technology. Featuring SRAM-based architecture, the chip dramatically reduces latency in sequential reasoning tasks. Nebius is the first cloud provider to integrate this hardware by December 2026. The launch coincides with Nvidia's upcoming earnings report, highlighting its ambitions in AI inference.