Reuters reports that OpenAI is developing its first custom AI chip, which is set to be completed by the end of this year. The company, in a collaboration with Broadcom and TSMC since mid-2024, will send the chip to TSMC for production testing at a 3-nanometer level.
According to Reuters, the OpenAI chip is expected to be ready in a few months, with mass production targeting the year 2026. The team lead designing the chip is Richard Ho, who previously worked at Google.
The goal of the OpenAI chip is to reduce dependence on NVIDIA chips and increase OpenAI’s negotiating power with chip manufacturers. While developing custom AI chips is not a new concept, successes like Google TPU, Meta MTIA, and Microsoft Azure MAIA 100 have yet to be matched by Meta and Microsoft.
Source: Reuters
TLDR: OpenAI is on track to release its first custom AI chip this year, aiming to reduce reliance on NVIDIA and strengthen its position in the market with mass production expected by 2026.
Leave a Comment