AIAI Infrastructure

OpenAI in Talks to Buy Google AI Chips to Reduce Reliance on Nvidia

ChatGPT's parent company has been renting Google processors since 2025 and is now weighing a direct purchase, a move that exposes the scramble for compute capacity to train and run artificial intelligence models

Curated by: Agent•Reviewed by: Marcelo Lourenço
Published on: Oct 29, 2026 03:35•Updated 53 minutes ago
OpenAI in Talks to Buy Google AI Chips to Reduce Reliance on Nvidia
OpenAI negocia compra de chips de IA do Google para reduzir dependência da NvidiaArte visual por Agente (Cloudflare Flux-1-Schnell)

OpenAI is evaluating the purchase of artificial intelligence chips manufactured by Google, according to reporting by American outlet The Information. The company behind ChatGPT has been renting processors from the search giant since 2025, and a possible purchase would represent a step beyond that commercial relationship.

So far, no details have been disclosed about the deal's value, the number of processors involved or the timeline for completing the purchase. Contacted by Reuters, Google did not respond to a request for comment, and OpenAI declined to comment on the matter.

The mechanics behind the dispute

Google's AI chips — TPUs (Tensor Processing Units) — are designed to accelerate the training and execution of machine learning models. Infrastructure is the central bottleneck of the artificial intelligence industry: training foundation models requires thousands of accelerators running in a cluster for weeks or months.

OpenAI has historically depended on Nvidia for this supply. Diversifying its processor base reduces the risk of shortages and expands the company's capacity for growth, as it faces direct competition from Google itself in the generative models market.

Rivalry and interdependence

The move highlights how competition in the AI sector does not prevent commercial relationships between companies vying for the same market. Google develops its own models and competes with OpenAI in products and services, but its processors can meet the infrastructure needs of other companies.

The information disclosed does not clarify whether OpenAI intends to replace part of the equipment it currently uses or merely expand its processing options. The decision to buy rather than rent would signal a long-term commitment to Google's platform.

What is at stake

The race for AI infrastructure has become the industry's silent battlefield. Whoever controls processing capacity controls the pace of new model releases — and, consequently, the competitive position in the market.

A possible purchase of Google chips by OpenAI would reinforce the search giant's position as an infrastructure supplier to rivals, a business model it already practices with other companies. For OpenAI, the deal would represent another layer of diversification at a time of accelerated expansion of its operations.

Share Story

Primary Source