OpenAI wins right to work with Microsoft rivals

Microsoft and OpenAI have significantly restructured their multi-billion-dollar partnership, allowing the AI company to offer its products through rival cloud providers including Amazon and Google while maintaining Microsoft as the primary cloud partner through 2032. The updated agreement, announced Monday, grants OpenAI the freedom to provide its AI models to any cloud provider and give API access to U.S. government national security clients regardless of their cloud infrastructure. However, Microsoft retains a non-exclusive license to OpenAI's intellectual property and will continue having its products launched first on Azure. On the business side, Microsoft will no longer pay revenue share to OpenAI, while OpenAI will still make revenue share payments to Microsoft until 2030, now subject to a total cap. Additionally, if Microsoft uses OpenAI technology to develop artificial general intelligence that surpasses human intelligence, the resulting models will face limits on computational capacity. The changes mark a significant shift in a relationship that has seen Microsoft invest more than $13 billion in OpenAI since 2019, though tensions reportedly grew in recent months over concerns the partnership limited OpenAI's ability to expand. An internal April memo from OpenAI's Chief Revenue Officer Denise Dresser indicated the alliance had restricted the company's reach. The timing coincides with OpenAI's growing partnerships with competitors, including a reported $50 billion multi-year strategic investment with Amazon announced in February to accelerate global generative AI innovation, on top of an earlier $38 billion infrastructure deal. Following the announcement, Microsoft's shares on Wall Street dipped 0.3 percent. In a separate regulatory development, the European Commission issued guidance to Alphabet's Google on Monday regarding how the tech giant must help online search rivals and AI developers access its services under the Digital Markets Act. Regulators identified that Google currently reserves key Android operating system capabilities exclusively for its Gemini AI service on smartphones and tablets. The proposed measures would require Google to ensure competing AI services can effectively interact with Android applications, such as sending emails through a user's preferred app, ordering food, or sharing photos. Google criticized the EU proposal, warning it would "strip away" device makers' autonomy, mandate access to sensitive hardware, and undermine privacy and security protections. Third parties have until May 13 to provide feedback before the Commission issues a final decision by the end of July on DMA compliance, with potential fines reaching up to 10 percent of annual global sales for violations.