OpenAI and Google Reach Agreement to Distribute Models to Google Cloud
OpenAI has reached a distribution agreement with Google Cloud to provide cutting-edge AI models, including GPT-6 Astra and Codex, to Google Cloud in the coming weeks.
Enterprise customers can purchase these models through the Google Cloud Marketplace, with related expenses counting towards existing cloud service spending commitments. The models will enter Google Cloud's Model Garden and support regulated industry clients in building agent-based workflows using Google's access control and data governance tools and infrastructure.
Google Cloud CEO Thomas Kurian stated that this move provides customers with a broader selection of models. OpenAI noted that this makes its strongest models and Codex available in customers' existing build environments. General availability is planned for the coming weeks.
Following the end of its exclusive distribution arrangement with Microsoft in April, OpenAI has expanded its distribution to AWS. Microsoft remains OpenAI's primary cloud partner.
This move allows OpenAI models to cover three major U.S. cloud platforms, deepening collaboration with competitor Google. The two companies previously collaborated on computing capacity and are now shifting to model distribution.
The market views this as the establishment of a multi-cloud distribution path, shifting funding from single cloud binding to multi-platform procurement, benefiting enterprise customers with model choices while putting pressure on traditional distribution models that rely on exclusive channels.
Source: Public Information
ABAB AI Insight
OpenAI previously maintained an exclusive distribution arrangement with Microsoft, which ended in April 2026, after which it expanded to AWS, continuing the shift from single cloud binding to multi-platform distribution.
In terms of capital pathways, OpenAI will sell models and Codex as SKUs through the Google Cloud Marketplace, shifting resources from direct APIs to cloud vendor channels, strategically leveraging existing cloud spending commitments to reduce procurement friction and expand reach.
Similar to other AI labs distributing through cloud platforms, the current phase is one of expanding models from exclusive channels to multi-cloud shelves, with cloud vendors becoming the primary distribution entry points for cutting-edge models.
This essentially represents a restructuring of the industry chain and a transfer of pricing power: models become deductible goods within cloud budgets, with mechanisms in place for cloud platforms to control procurement processes and commitment amounts, concentrating resources on platforms that provide both their own models and third-party models.
ABAB News · Cognitive Law
After the end of exclusive distribution, models enter multi-cloud shelves.
Cloud commitment amounts can offset model purchases, reducing friction through channels.
Competitors collaborate on distribution, concentrating platform choice with cloud vendors.