On the evening of April 27,

According to the revised agreement, Microsoft will continue to maintain its position as OpenAI's primary cloud partner, and OpenAI's products will continue to be prioritized on Microsoft Azure. However, OpenAI is now authorized to provide its full range of services to global customers through any cloud service provider. In terms of intellectual property, the licensing validity period for Microsoft's authorization of OpenAI models and related products will extend until 2032. At the same time, the policy of OpenAI paying revenue-sharing to Microsoft will continue until 2030, although the sharing ratio remains unchanged and has an upper limit. However, this payment will no longer be affected by the progress of OpenAI's technological iterations.
Although the financial and licensing terms have been decoupled, both parties still emphasized their deep collaboration on underlying computing power and cutting-edge research. Microsoft will continue to participate in OpenAI's long-term development as an important shareholder. Both sides will jointly promote the construction of petawatt-level data centers, co-develop next-generation AI chips, and explore the use of artificial intelligence to upgrade cybersecurity protection.
Amid the current industry context where open-source models are accelerating their iterations and AI companies are emerging intensively, the flexible adjustment of the relationship between Microsoft and OpenAI not only grants OpenAI greater market autonomy but also reflects the industry trend of model developers and cloud giants seeking a new balance between resource binding and commercial independence.
