May 27, 2025Published by Sean Kim on May 27, 2025Categories Industry NewsGroq LPU Inference at 500+ Tokens/Second: The Chip That Could Dethrone NVIDIA in Real-Time AI625 tokens per second. That’s not a typo, and it’s not a synthetic benchmark from a lab nobody’s heard of. That’s Groq’s LPU running Meta’s Llama […]