OpenAI will start using AMD chips and may create its own AI hardware in 2026

OpenAI is reportedly working with Broadcom to develop up-to-date custom silicon designed to handle enormous AI workloads for inference and secure manufacturing capacity at TSMC, according to sources they talk to Reuters. OpenAI has reportedly built a chip development team of around 20 people, including lead engineers who previously worked on Google Tensor processors for artificial intelligence.

Nevertheless, according to the current schedule, production of specially designed equipment may not begin until 2026.

Information in July reported that OpenAI was in talks with Broadcom and other semiconductor designers about developing its own AI chip, and earlier this year Bloomberg announced that OpenAI is working on building its own foundry network, but according to Reutersthese plans were shelved for cost and time reasons.

The reported strategy puts OpenAI on a similar path to other tech companies trying to manage costs and access to AI server hardware with custom chip designs. However, Google, Microsoft and Amazon are already several generations older in their efforts, and OpenAI may need much more resources to become a real competitor.