OpenAI and Microsoft Part Ways, Joins AWS to Expand Multi-Cloud Strategy

OpenAI officially announced a restructured partnership with Microsoft on April 29, 2026, ending six years of exclusive cloud collaboration. Under the new agreement, OpenAI can now deploy and sell its AI models on Amazon Web Services and Google Cloud without needing Microsoft’s approval. This shift marks OpenAI’s move from single-cloud reliance to a multi-cloud strategy, designed to broaden its enterprise market coverage.

In exchange, Microsoft will no longer pay OpenAI a revenue share from Azure cloud services. Microsoft will retain all revenue generated from selling OpenAI models through Azure going forward. OpenAI, in turn, gives up the steady income it once received from Microsoft cloud sales in exchange for full operational independence. The two parties also removed previous clauses that would have altered the business relationship upon the achievement of artificial general intelligence, clearing the way for the new structure.

Amazon Web Services promptly issued a statement: “This marks the starting point for deeper collaboration between AWS and OpenAI.” The two sides will bring the latest advancements to the Amazon Bedrock platform and launch the Codex tool, helping users build production-grade AI agents using OpenAI’s frontier models. According to Axios, Google Cloud is also evaluating the new agreement and may join the partnership in the future.

Commentary: This partnership restructuring reshapes the competitive landscape of AI cloud services. By breaking free from an exclusive tie, OpenAI can now reach a more diverse customer base more flexibly. Microsoft, while losing the revenue share, gains higher own-cloud revenue. The rapid formation of a multi-cloud ecosystem could become a core trend for enterprise AI deployment in 2026.