OpenAI reportedly eyes Google TPU purchases as its chip supply broadens

OpenAI plans to purchase Google’s tensor processing units (TPUs), according to a TechRadar report citing The Information. The proposed purchase would broaden the ChatGPT developer’s processor supply as it seeks more computing resources for its AI services.

The report does not disclose financial terms, chip quantities or delivery schedules. Those omissions leave the scale and timing of the proposed purchase unclear, including how much capacity Google’s hardware might contribute alongside OpenAI’s existing Nvidia processors.

A purchasing plan with an existing relationship

Google’s chips would not be entirely unfamiliar territory for OpenAI. According to TechRadar, OpenAI began renting Google’s AI chips in 2025. The reported purchasing plan would therefore extend a computing relationship that was already underway.

TechRadar also reports that OpenAI recently announced plans to make its advanced AI models available through Google Cloud. Together, these arrangements illustrate how companies competing in AI development can also become commercial partners in infrastructure and distribution.

Alphabet has expanded access to processors previously used mainly within Google’s own infrastructure, supplying selected external customers. OpenAI, meanwhile, remains a major Nvidia customer, using its GPUs to develop and operate AI systems.

Custom processors remain part of the plan

External suppliers are only part of OpenAI’s hardware strategy. TechRadar says the company announced in June 2026 that it had designed an AI chip with Broadcom, with deployment intended before the end of 2026. That remains a stated target, rather than confirmation that deployment has happened.

For readers following Nvidia and PC technology, the relevant distinction is the market involved: this report concerns processors for AI infrastructure. It provides no announcement about consumer graphics cards, gaming performance or console hardware.