'Buying from its biggest rival' — OpenAI turns to Google for TPUs to break free from Nvidia's GPU monopoly | Daily Reports Online

OpenAI plans to buy Google’s AI chips as computing demand rises
Google’s TPUs are attracting customers seeking alternatives to Nvidia’s GPUs
OpenAI plans to deploy its custom-designed AI chip before 2027

OpenAI plans to purchase artificial intelligence chips from Google as the ChatGPT developer expands its computing resources beyond its established reliance on Nvidia processors.

According to The Information, the proposed purchase involves Google’s tensor processing units (TPUs), although financial terms, quantities, and delivery schedules remain undisclosed.

The reported agreement follows growing demand for AI computing infrastructure, with technology companies seeking additional processors to support increasingly expensive model training and deployment.

Google attracts OpenAI as AI chip demand increases

Google’s TPUs have attracted customers seeking alternatives to Nvidia’s GPUs, which remain widely used for training and operating artificial intelligence models.

Alphabet, Google’s parent company, has expanded commercial sales of its processors, allowing selected external customers to obtain chips previously used mainly within Google’s own computing infrastructure.

OpenAI already ranks among Nvidia’s major customers, purchasing GPUs for developing and running its AI systems, including the models that power ChatGPT and other services.

However, the reported Google purchase indicates that OpenAI intends to obtain computing capacity from another supplier rather than depend exclusively on Nvidia’s hardware.

In 2025, OpenAI started renting Google’s AI chips to support its products, indicating that the relationship predates the latest purchasing plans.

The proposed transaction also follows OpenAI’s recent announcement that its advanced AI models would become available through Google Cloud, adding another commercial connection between the competing companies.

Despite competing in AI development and other technology markets, Google and OpenAI have increasingly entered arrangements involving computing resources and access to their respective products.

OpenAI develops custom chips alongside external purchases

OpenAI has also pursued its own processor development programme, announcing in June 2026 that it had designed an AI chip with semiconductor company Broadcom.

The company said it intended to deploy that processor before the end of 2026, forming part of a multi-generation effort to develop specialised computing hardware.

The custom chip programme, alongside purchases from Nvidia and the reported Google deal, indicates that OpenAI is pursuing several sources of computing capacity.

Google’s growing chip business could benefit from this demand, particularly as Alphabet seeks commercial returns from its substantial investments in AI infrastructure.

Alphabet’s 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.

She added that most revenue associated with those agreements was expected to be recognised during 2027, suggesting that the financial contribution would develop over time.

Alphabet, Amazon, and Microsoft are collectively expected to spend as much as $600 billion on AI infrastructure during 2026, according to the report.

Such expenditure has attracted investor scrutiny, with questions emerging about when investments in computing capacity will generate sufficient financial returns.