NVIDIA Secret Inference Chip with Groq LPU Technology

NVIDIA is building a new inference processor integrating Groq's Language Processing Unit technology (acquired December 2025). The chip uses on-chip SRAM instead of HBM, delivering up to 80 TB/s memory bandwidth (~10x H100). OpenAI committed to 3 GW of dedicated inference capacity as lead customer. Debut at GTC 2026 (March 16-19). Meanwhile, Jensen Huang told Morgan Stanley that a "$100 billion investment in OpenAI is probably not in the cards" — notable distancing during the backlash. (SiliconANGLE)