Microsoft and OpenAI have announced a collaboration on a new data center project. One of the main objectives of the project is to develop a giant AI supercomputer known as Stargate. According to The Information, Microsoft is expected to pay a significant portion of the cost of the Stargate supercomputer, which is estimated to be over $100 billion. This figure is staggering, especially when compared to the cost of other large data center projects currently in operation.
The joint project between OpenAI and Microsoft could be 100 times more costly than some of the largest data center projects in existence. This estimate comes from sources familiar with and involved in discussions about the project, including a person who spoke with OpenAI CEO Sam Altman and another person who had access to Microsoft’s initial cost analysis.
It has been reported that the Stargate supercomputer will be located in the United States and will become the largest of many installations that Microsoft and OpenAI are planning to build in the coming years. The supercomputer is expected to go into operation after 2028, and it marks the fifth and final phase of Microsoft’s current supercomputer and OpenAI projects. The two companies are currently in phase three, and in phase four, Microsoft is researching a new, smaller supercomputer that promises to launch in 2026.
It is noteworthy that the majority of the costs in this five-phase plan of the Microsoft and OpenAI joint venture stem from the need to purchase AI chips. This means that Nvidia is likely to benefit significantly from this project with a budget of over $100 billion.
Although it is unclear what Stargate is capable of, its development is not surprising given the current demand for AI models and the continued growth of AI. OpenAI and Microsoft are continuing to push the boundaries of AI, and with GPT-5 on the horizon, large data centers such as Stargate are becoming increasingly necessary.