Microsoft and OpenAI Amend Partnership Deal, Allowing OpenAI More Cloud Flexibility

Microsoft and OpenAI announced an amended agreement on April 27, 2026, significantly restructuring their landmark partnership. The revised terms strip away Microsoft's previous revenue share payments to OpenAI, effectively ending a financial arrangement that had been part of earlier deals. More importantly, Microsoft's exclusive license to OpenAI's intellectual property becomes non-exclusive, meaning OpenAI can now license its technology to other major cloud providers. OpenAI gains the right to use alternative cloud infrastructure if Microsoft Azure cannot or chooses not to support the specialized hardware and capabilities needed for training and running advanced AI models.

This move signals a strategic shift for both companies. For OpenAI, the new cloud flexibility solves a growing bottleneck: as models scale to trillion-parameter sizes, relying solely on Azure for compute capacity risks throttling development. The amendment caps OpenAI's revenue share payments to Microsoft until 2030, offering predictable financial terms. For Microsoft, relinquishing exclusivity likely protects against antitrust concerns while retaining its role as OpenAI's primary investor. The deal allows OpenAI to potentially tap AWS or Google Cloud for peak workloads, giving it negotiating leverage and ensuring model training isn't hindered by capacity constraints.