OpenAI Plans to Use Google TPUs, Challenging Nvidia's AI Chip Dominance

On October 27, Google announced that Oracle will become the first major cloud provider to rent TPUs to customers. The chips are being purchased directly by Oracle, with the end user being a major customer shared by both companies, according to sources familiar with the matter.

Separately, OpenAI is reportedly planning to procure Google's tensor processing units (TPUs). Citing people familiar with the matter, The Information reported that OpenAI has posted a job listing seeking an engineer to "lead" optimization of its most advanced frontier models on TPUs, indicating the effort is moving into a substantive phase.

Google has developed TPUs for more than a decade and has rented them to external customers through its own cloud infrastructure, including users such as Anthropic and Apple. More recently, Google has begun selling TPUs directly to other data center operators, and the Oracle agreement marks its latest expansion into cloud-provider channels.

OpenAI has a close relationship with Nvidia. In August 2026, Nvidia agreed to provide more than $100 billion in credit support for a large data center project OpenAI is planning in Ohio, which will fully use Nvidia's chip architecture. Nvidia CEO Jensen Huang said at the time that OpenAI had committed to spending about $600 billion on Nvidia servers by 2030. Still, OpenAI is also seeking to diversify its chip supply and reduce its reliance on Nvidia.