OpenAI plans to purchase artificial intelligence chips from Google as it expands its computing resources beyond its established reliance on Nvidia processors. The reported deal involves Google's tensor processing units (TPUs), although its financial terms, chip quantities and delivery schedule have not been disclosed, according to The Information.
The move comes as demand for AI infrastructure continues to grow and companies look for additional processors to support increasingly costly model training and deployment. Nvidia GPUs remain widely used for developing and operating AI systems, and OpenAI is already one of the chipmaker's major customers through its work on ChatGPT and other services.
Google's TPUs have attracted customers seeking alternatives to Nvidia's hardware. Alphabet, Google's parent company, has expanded commercial sales of the processors, which were previously used mainly for Google's own computing operations.
OpenAI began renting Google's AI chips in 2025 to support its products, indicating that the relationship predates the reported purchasing plans. The two companies have also established another commercial link after OpenAI announced that its advanced AI models would become available through Google Cloud.
At the same time, OpenAI is developing custom hardware. In June 2026, the company announced that it had designed an AI chip with semiconductor manufacturer Broadcom. OpenAI said it intended to deploy the processor before the end of 2026 as part of a multigeneration effort to create specialised computing hardware.
The combination of Nvidia purchases, Google's reported TPU deal and its own chip programme suggests that OpenAI is pursuing several sources of computing capacity. The additional demand could also support Google's growing chip business as Alphabet seeks commercial returns from its investments in AI infrastructure.
Alphabet finance chief Anat Ashkenazi previously said the company expected to recognise a small portion of revenue from TPU sales agreements by the end of 2026, with most of the related revenue expected during 2027.
The report also said Alphabet, Amazon and Microsoft could collectively spend as much as $600 billion on AI infrastructure during 2026. That level of expenditure has drawn investor scrutiny over when investments in computing capacity will produce sufficient financial returns.
Sources: TechRadar.