The next phase of the Microsoft OpenAI partnership

The next phase of the Microsoft OpenAI partnership

Summary (EN)

OpenAI announced an amended agreement with Microsoft that significantly reshapes the economic and infrastructure terms of one of the most consequential partnerships in the AI industry. The updated arrangement keeps Microsoft as OpenAI’s primary cloud partner and preserves Azure-first shipping for OpenAI products unless Microsoft cannot or declines to support the required capabilities. At the same time, OpenAI says it can now serve its products to customers across any cloud provider, introducing more commercial flexibility than the previous setup implied. The amended terms also make Microsoft’s license to OpenAI intellectual property non-exclusive through 2032, remove Microsoft’s obligation to pay revenue share to OpenAI, and keep OpenAI’s revenue-share payments to Microsoft in place through 2030 subject to the same percentage and a total cap. Microsoft remains a major shareholder and continues to participate directly in OpenAI’s growth. OpenAI frames the change as a move toward greater predictability and simpler cooperation while both companies continue working together on large-scale data center capacity, next-generation silicon, cybersecurity applications, and global AI deployment. The announcement matters because it clarifies the balance between strategic interdependence and competitive freedom. Microsoft remains central to OpenAI’s infrastructure and commercialization, but OpenAI is explicitly preserving room to sell and operate across multiple cloud environments. For the broader market, the agreement is an industry-structure signal: the leading lab-cloud alliance is not dissolving, but it is being restructured around more modular economics, less exclusivity, and a wider distribution model for frontier AI services.

Summary (ZH)

OpenAI 宣布与微软修订合作协议，这一变化重新定义了 AI 行业中最重要合作关系之一的商业和基础设施安排。根据新条款，微软仍将是 OpenAI 的主要云合作伙伴，OpenAI 产品原则上仍优先在 Azure 上首发，除非微软无法或选择不支持所需能力。与此同时，OpenAI 明确表示，它现在可以把产品提供给任何云平台上的客户，这意味着相比过去更具商业灵活性。新协议还规定，微软对 OpenAI 模型与产品知识产权的许可将持续至 2032 年，但从独占改为非独占；微软不再向 OpenAI 支付收入分成；而 OpenAI 向微软支付的收入分成将继续到 2030 年，比例不变但设有总上限。微软仍是重要股东，并继续直接参与 OpenAI 的增长。OpenAI 将此次修订描述为在保持合作雄心的同时，提升确定性并简化双方协作机制。双方未来仍将共同推进大规模数据中心建设、下一代芯片协作、网络安全应用和 AI 全球部署。此举的重要性在于，它更清楚地划定了“深度绑定”与“竞争自由”之间的边界。微软仍然是 OpenAI 基础设施和商业化路径中的关键角色，但 OpenAI 也公开保留了跨云服务和更广分发的空间。对整个市场而言，这说明头部实验室与头部云厂商的联盟并未瓦解，而是在向更模块化的经济结构、更低排他性以及更广覆盖的前沿 AI 服务分发模式演化。

Source

https://openai.com/index/next-phase-of-microsoft-partnership/