The Financial Times reported yesterday, citing informed sources, that OpenAI is set to produce its own AI chip for the first time next year in partnership with American semiconductor company Broadcom.

The report added, citing a source familiar with the project, that OpenAI intends to use the chip internally rather than making it available to external customers.

Reuters has not been able to verify the report’s accuracy yet.

Neither company responded to Reuters’ requests for comment after business hours.

Last year, Reuters reported that OpenAI was collaborating with Broadcom and Taiwan Semiconductor Manufacturing Company (TSMC) to develop its first in-house chip to run its AI systems, integrating chips from AMD and Nvidia to meet growing infrastructure demand.

At that time, OpenAI was exploring options to diversify chip sources and reduce costs.

Reuters reported in February that OpenAI is seeking to reduce its reliance on Nvidia for chip supplies by developing its first generation of AI chips.

Broadcom CEO Hock Tan said yesterday that the company expects a significant improvement in AI revenue growth for fiscal year 2026 after receiving orders worth more than $10 billion for AI infrastructure from a new, unnamed customer.

OpenAI’s move follows efforts by Google, Amazon, and Meta, which have developed custom chips to handle AI workloads amid increasing demand for technological capabilities to train and operate AI models.