OpenAI Turns to Google for More GPUs

OpenAI has struck a deal with Google Cloud to meet its rising computing demands, sources told Reuters, marking a rare collaboration between AI rivals.

Finalised in May after months of talks, the agreement provides OpenAI with additional cloud infrastructure beyond Microsoft Azure, its primary provider to date.

The move highlights the immense computing power needed to train and run AI models and reflects OpenAI’s growing efforts to diversify resources.

OpenAI previously partnered with Oracle and SoftBank on its $500 billion Stargate project and signed multibillion-dollar compute deals with CoreWeave. It is also working on its own AI chips.

For Google Cloud, which contributed $43 billion to Alphabet’s 2024 revenue, the OpenAI deal is a win amid competition from AWS and Azure.

It adds to a customer list that includes Apple and AI startups like Anthropic.

However, the partnership complicates Google's internal resource management, especially as OpenAI’s ChatGPT competes directly with Google Search and DeepMind’s models.

While analysts called the deal “surprising,” they noted the strategic value in supplying compute—even to rivals.

Previously, OpenAI CEO Sam Altman, taking a dig at Google, stated, “I don’t do Google searches anymore.” At a U.S. Senate hearing, Altman shared that he personally prefers ChatGPT for searching.

However, he acknowledged Google remains a “ferocious competitor” with a strong AI team and is rapidly integrating AI into search.

OpenAI also has made a habit of timing its major announcements just days before Google’s flagship events, a strategic move seen by many as an attempt to steal the spotlight and shape the AI narrative ahead of its biggest rival.