OpenAI plans to buy Google’s artificial intelligence chips, according to a 27 October report citing a person familiar with the matter.
The proposed purchase would deepen OpenAI’s ties with Alphabet’s Google. It comes after OpenAI began renting Google’s AI chips in June 2025, Reuters reported, and after it said last week that its frontier AI models would be available through Google Cloud.
The arrangement has a faintly unusual quality. OpenAI and Google compete in AI products and search, while OpenAI is also already among Nvidia’s largest buyers of graphics processing units, or GPUs.
Google’s Tensor Processing Units, known as TPUs, have become an alternative to Nvidia GPUs for training AI models and running inference, the process of generating responses from a trained model. Google began selling TPUs directly to some customers in 2026. It also announced plans to sell them to select customers for use in their own data centres, making the chips available beyond Google Cloud.
Demand for Google’s chips has been helped by agreements with customers including Anthropic, which competes with OpenAI.
OpenAI is building another route to computing capacity. In June, it unveiled its first custom AI chip, designed with Broadcom, and said it expected to deploy the processor by the end of this year as part of a multi-generation chip programme.
The wider spending race is substantial: Amazon, Microsoft and Alphabet plan to spend as much as $600 billion combined this year on AI infrastructure. For Alphabet, chip sales offer a potential new revenue line alongside its cloud operation. It expects to recognise a small percentage of revenue from TPU sales agreements by the end of this year, with the vast majority expected in 2027.
Alphabet is due to report results on 28 October. Investors are watching whether its heavy AI spending continues to generate growth in the cloud business.
Comments (0)