Microsoft (MSFT, Financial)-backed OpenAI is moving forward with plans to design its own AI chips and is working with Taiwan Semiconductor Manufacturing (TSM, Financial) to bring them to production, according to a Reuters report.
The ChatGPT creator is expected to finalize chip designs within the next few months before sending them to TSMC for fabrication, with mass production targeted for 2026. The chips will reportedly use TSMC's advanced 3-nanometer process technology, helping OpenAI reduce its reliance on Nvidia (NVDA), the dominant provider of AI hardware.
Last year, reports surfaced that OpenAI had scrapped its broader foundry ambitions and instead partnered with Broadcom (AVGO, Financial) and TSMC for its custom chip efforts. In the meantime, OpenAI will continue using chips from Nvidia and Advanced Micro Devices (AMD, Financial) to power its AI models. On Sunday, the company aired its first-ever Super Bowl ad, highlighting its pursuit of Artificial General Intelligence.
