OpenAI is reportedly preparing to launch its first in-house artificial intelligence chip in 2026 through a partnership with Broadcom, with plans to deploy the silicon internally rather than sell it to outside customers, according to a Financial Times report cited by Reuters. Reuters said it could not immediately verify the report, and noted that OpenAI and Broadcom did not respond to requests for comment outside regular business hours.
Partnership points to internal deployment
The Financial Times report, relayed by Reuters, said OpenAI’s first chip would be produced with Broadcom and used inside the company to power its own AI systems. The approach aligns with OpenAI’s need for substantial computing power to train and run generative AI models and comes amid industry-wide efforts to secure chip supply and manage infrastructure costs.
Reuters previously reported that OpenAI had been working with Broadcom and Taiwan Semiconductor Manufacturing Co to develop its first in-house chip, while continuing to incorporate AMD chips alongside Nvidia hardware to meet rising compute demand. At the time, the company was examining options to diversify chip supply and reduce costs, and was finalizing a first-generation design to send for fabrication at TSMC, Reuters reported.
Broadcom signals growing AI pipeline
Broadcom CEO Hock Tan said the company expects artificial intelligence revenue growth for fiscal 2026 to “improve significantly,” following more than $10 billion in AI infrastructure orders from a new, unnamed customer, according to Reuters. Tan said the prospect placed a firm order last quarter, qualifying as a customer, and earlier this year had pointed to four potential customers “deeply engaged” on custom chips in addition to three existing large clients.
Custom silicon trend spreads across big tech
Reuters noted that OpenAI’s move would follow similar efforts by Google, Amazon and Meta, which have built custom chips to handle AI workloads as demand for computing resources increases. The reported plan underscores how AI developers are pursuing tailored silicon strategies to support training and inference at scale.
Reuters emphasized that it could not immediately verify the Financial Times’ report regarding OpenAI’s 2026 timeline and internal-use focus. The outlet added that OpenAI and Broadcom did not immediately respond to its requests for comment. Prior Reuters coverage detailed OpenAI’s push to reduce reliance on Nvidia by developing its first in-house AI silicon and working with manufacturing partners.
Reporting was credited to Disha Mishra in Bengaluru, with editing by Alan Barona and Sherry Jacob-Phillips, as carried by Reuters.
