According to Reuters, OpenAI is collaborating with TSMC and Broadcom to develop internal AI chips and has begun using AMD and Nvidia chips for AI training. OpenAI plans to launch its first AI chip in 2026.
OpenAI has been working with Broadcom for several months to develop AI chips for running models, which could debut as early as 2026. Meanwhile, OpenAI plans to use AMD chips for model training via Microsoft's Azure cloud platform.
Previously, OpenAI relied almost entirely on Nvidia GPUs for training, but chip shortages, delays, and high training costs have prompted OpenAI to explore alternative solutions. OpenAI has abandoned plans to establish a network of chip manufacturing plants, focusing instead on internal chip design.