Microsoft and OpenAI are collaborating on an ambitious project to build a cutting-edge data center called “Stargate,” which is anticipated to launch in 2028.
The initiative, estimated to cost up to $100 billion, aims to address the soaring demand for AI-capable data centers fueled by the rapid adoption of generative AI technology.
According to reports, Microsoft is likely to foot the bill for this colossal undertaking, which dwarfs the costs of existing data centers by a factor of 100. The proposed supercomputer, set to be the largest in a series of such projects over the next six years, is expected to be based in the U.S.
The plan, spearheaded by OpenAI’s CEO, Sam Altman, and Microsoft, is structured into five phases, with Stargate representing the final phase. Preceding this, a smaller supercomputer is in the works for OpenAI, scheduled for launch around 2026. The substantial costs associated with the project primarily involve procuring specialized AI chips, which are currently priced at premium rates, such as Nvidia’s “Blackwell” B200 chip, priced between $30,000 and $40,000.
Microsoft has also developed its own custom computing chips, announced in November of the previous year, and intends for the new project to be compatible with chips from various suppliers. Although Microsoft declined to comment directly on the Stargate supercomputer, a spokesperson emphasized their commitment to advancing AI capabilities through infrastructure innovation.
The projected expenses for the project could surpass $115 billion, significantly exceeding Microsoft’s capital spending on servers and infrastructure from the previous year
“We are always planning for the next generation of infrastructure innovations needed to continue pushing the frontier of AI capability,” a Microsoft spokesperson said in an emailed statement to Reuters. The spokesperson did not comment directly on the report about the planned launch of the Stargate supercomputer.
