OpenAI Shifts ChatGPT AI Chips From Nvidia to Google Cloud

OpenAI has begun renting Google's tensor processing units (TPUs) to power ChatGPT and other AI products, marking its first significant use of non-Nvidia chips. This shift, moving away from Microsoft data centers and Nvidia GPUs, aims to reduce the high costs associated with inference computing. However, Google is limiting OpenAI's access to its most advanced TPUs to maintain a competitive edge, reserving those for internal projects like its Gemini language model. The partnership highlights a strategic move by OpenAI to diversify its AI chip supply amid growing demand, while Google expands its TPU customer base to include major tech players such as Apple. This development could disrupt the AI hardware market by challenging Nvidia's dominance and encouraging more cross-company collaborations in the AI sector. Overall, it reflects evolving industry dynamics where cost, availability, and technological innovation drive new infrastructure strategies for AI companies.