OpenAI is putting its AI models and its Codex coding tools on Google Cloud. With that move, customers of all three major US cloud providers can now reach OpenAI’s technology from their existing platform.
What OpenAI is offering on Google Cloud
Google Cloud customers will get access to OpenAI’s frontier models and to Codex, the company’s coding tool. Developers inside the Google ecosystem no longer need a separate route to use OpenAI’s products.
This rollout follows a similar one on Amazon’s side. OpenAI’s models became available through Amazon Bedrock, AWS’s managed AI service, on April 28, 2026.
Microsoft Azure has long been OpenAI’s home base. With Google Cloud added, OpenAI now covers the full trio of dominant US cloud platforms.
How a Microsoft-first company got here
For years, OpenAI leaned heavily on an exclusive relationship with Microsoft for its cloud needs. Diversification became necessary as usage surged and OpenAI needed far more computing capacity than any single partner could easily supply.
The Google relationship started on the infrastructure side. OpenAI finalized a deal to use Google Cloud infrastructure in May 2025, after getting past blockages tied to its Microsoft exclusivity.
Two months later, in July 2025, Google Cloud officially joined OpenAI’s list of suppliers. It sat alongside Microsoft, Oracle and CoreWeave.
OpenAI also began renting Google’s TPUs, the custom chips Google designs for AI workloads. That marked a shift away from the Nvidia chips that have powered much of the AI boom.
Underneath all of this is revenue. OpenAI reported an annualized revenue run rate of $10 billion as of June 2025. A business growing at that scale needs a lot of compute, and it needs it from more than one landlord.
The latest announcement flips the direction of the relationship. Until now, OpenAI was mostly a Google Cloud customer buying capacity. Now Google Cloud’s own customers can buy OpenAI’s products.
What this means for the enterprise AI race
For Microsoft, the picture is more complicated. Its early, close partnership with OpenAI was a key selling point for Azure. That edge narrows as OpenAI’s tools appear on rival clouds.
The TPU detail deserves attention too. OpenAI’s decision to rent Google’s chips shows the largest AI developers are willing to look beyond Nvidia when capacity and cost demand it.
Disclosure: This article was edited by Diego Almada Lopez. For more information on how we create and review content, see our Editorial Policy.