Microsoft Relaxes Data Center Control Over OpenAI in $500B Oracle-SoftBank Venture
The core elements of their partnership remain intact through the duration of their contract, which extends until 2030.

Microsoft has adjusted its partnership terms with OpenAI, allowing the latter to expand its data centre capabilities while still retaining exclusive rights to OpenAI’s application programming interface (API).
This change comes amid the announcement of a significant joint venture named “Stargate,” involving Oracle and Japan’s SoftBank Group, aimed at building up to $500 billion in new AI data centres in the United States.
Microsoft, which has held exclusive rights to build computing infrastructure for OpenAI since 2019, has now approved OpenAI to develop additional capacity primarily for research and model training. This shift enables OpenAI to collaborate with Oracle on infrastructure projects.
The Stargate initiative is structured as a new entity where OpenAI will have an equity stake, governance rights, and operational control. It will feature a separate board and CEO, with other investors including the UAE firm MGX. Microsoft, while not an equity funder in this venture, will serve as a technology partner alongside Nvidia and Arm.
Despite these changes, Microsoft maintains its exclusive right to offer OpenAI’s API, which is crucial for software developers and business customers seeking to utilize OpenAI’s services. This means that Oracle will not be able to host OpenAI’s main revenue-generating services.
Microsoft confirmed that it has established revenue-sharing agreements with OpenAI that benefit both parties. The core elements of their partnership remain intact through the duration of their contract, which extends until 2030.