Our portfolio founders Reiner Pope and Mike Gunter were interviewed by Joe Weisenthal and Tracy Alloway of Bloomberg's Odd Lots podcast. In this episode, they discussed how MatX is building the ultimate chip specifically for LLMs and how it could compete with the AI chip giant NVIDIA.
Reiner and Mark also discussed the process of designing GPUs, why hyperscale companies like Google and Meta don't build their own LLM-specific chips, how different MatX is from NVIDIA, the effect of LLM learning styles on chip design, the costs of doing business, and more.
Take a listen on Bloomberg's website, YouTube, or Spotify.
