TAIPEI (Taiwan News) — Industry sources said OpenAI plans to send its finalized chip design to TSMC in the coming months, Reuters reported Monday.

Sources said this will be the ChatGPT creator’s first in-house chip, marking a move away from its reliance on Nvidia. The process of sending a final chip design to a fab is called “taping out.”

TSMC is a pure-play foundry, meaning it does not design chips, but manufactures them for other companies. The Hsinchu-based company produces about 90% of the world’s advanced chips and is a major supplier to Apple and Nvidia.

According to Reuters, the update shows that OpenAI is serious about its goal to begin mass production at TSMC in 2026. Open AI head of hardware Richard Ho is leading the team, which has doubled to 40 in the past months.

The new chip, to be made with TSMC’s 3 nm process, features a systolic array architecture with high-bandwidth memory and extensive networking capabilities, sources explained. It will be used mainly for running AI models and will have a limited role in the company’s infrastructure.

There is no guarantee the silicon will function on the first tape out and a failure would require the company to diagnose the problem and repeat the tape out step, per Reuters.

The report added that Microsoft and Meta have also been making inroads to expand AI capabilities. Microsoft said it would invest US$80 billion (NT$2.6 trillion) in AI infrastructure this year, whereas Meta said it would spend US$60 billion next year.

However, rising costs and high demand for chips from a single supplier have led them to follow OpenAI’s steps and seek alternatives to Nvidia’s silicon.