OpenAI plans to buy Google's TPU chips and diversify beyond Nvidia
The reported purchase would turn two AI rivals into closer infrastructure partners, but neither the order size nor price has been disclosed.

OpenAI plans to buy Google’s AI chips, according to The Information, with Reuters and Bloomberg also reporting the move. The purchase would give one of Nvidia’s largest customers another source of accelerators as demand for AI computing keeps rising.
The chips are Google’s Tensor Processing Units, or TPUs, which the company designed for training and running large AI models. Google has historically used TPUs inside its own products and cloud platform, but it is increasingly trying to sell the hardware and related capacity to outside customers.
For OpenAI, adding Google chips would broaden a computing strategy that already depends heavily on Nvidia and includes work on other hardware options. It would also deepen an unusual relationship between rivals: OpenAI and Google compete in models, assistants and search, while also working together through cloud and infrastructure deals.
The practical significance is pressure on Nvidia’s dominance. A large OpenAI order could validate TPUs as a more direct alternative for frontier AI workloads and encourage other buyers to use multiple chip suppliers. Neither company has confirmed the report, and the size, price and delivery schedule of the potential order were not disclosed.



