OpenAI partners with Broadcom to launch first AI chip

OpenAI is developing its first artificial intelligence chip in partnership with semiconductor manufacturer Broadcom, with production scheduled for 2026.

The custom silicon will be designed for internal use rather than commercial sale, marking the ChatGPT maker's latest move to secure computing resources amid soaring AI infrastructure demands.

The collaboration continues OpenAI's previously reported efforts to diversify its chip supply and reduce costs.

TSMC and Broadcom collaboration with chatbot manufacturer was unveiled through news reports that tech giants are working on in-house chip development while simultaneously incorporating AMD chips alongside its primary Nvidia infrastructure.

American multinational designer company CEO Hock Tan hinted at the partnership: “The company had secured more than $10 billion in AI infrastructure orders from a new qualified customer without identifying the client.”

Tan had previously referenced four new potential customers deeply engaged in custom chip development alongside three existing major clients.

The move fits with strategies implemented by other tech giants including Google, Amazon, and Meta, all of which have developed custom chips to handle specialized AI workloads as demand for computational power surges.

Industry analysts states that custom chip development represents both a significant investment and potential long-term savings for AI companies facing constrained GPU supply chains and escalating operational costs.