OpenAI is partnering with Broadcom to develop custom silicon for AI workloads, securing production capacity with TSMC but not expecting to begin production until 2026. Meanwhile, OpenAI is using AMD’s MI300 chips in its Azure setup to support immediate needs.
This custom chip initiative aligns OpenAI with other tech giants like Google and Microsoft, though they are further along in similar efforts. OpenAI’s plans for its own foundries have been paused due to high costs and long timelines, and it may need substantial funding to compete effectively in the custom AI hardware space.
Read more – https://www.theverge.com/2024/10/29/24282843/openai-custom-hardware-amd-nvidia-ai-chips
