OpenAI is working with Broadcom and Taiwan Semiconductor Manufacturing Co. (TSMC) to develop a specialized AI inference chip.

These discussions are in the early stages and have been ongoing for about a year.

The focus of this collaboration is on creating a chip that can efficiently run software and handle user requests, rather than competing with Nvidia in the GPU market.

Analysts predict a growing demand for AI inference chips as more companies adopt complex AI models.

OpenAI has chosen not to comment on the discussions, while Broadcom and TSMC have not provided any response or confirmation.