NVIDIA acquired Groq and have already integrated their technology in their solution. It seems it can really increase inference performance per energy used. I do

NVIDIA acquired Groq and have already integrated their technology in their solution. It seems it can really increase inference performance per energy used. I don’t believe The benchmark compares Jalapenjo together with this solution. https://nvidianews.nvidia.com/news/nvidia-groq-3-lpx-now-in-full-production-with-world-class-speed-for-agentic-ai https://nvidianews.nvidia.com/news/nvidia-groq-3-lpx-now-in-...