Microsoft and OpenAI Unveil Ambitious US$100 Billion Data Center Project

Microsoft and OpenAI have embarked on an extensive data center project with plans that could see investments reaching up to a staggering US$100 billion ($153.4 billion). Central to this ambitious venture is the creation of an advanced artificial intelligence supercomputer named “Stargate.” According to reports from The Information, this state-of-the-art AI supercomputer is scheduled to be unveiled in 2028.

OpenAI has yet to provide a response to requests for comment from Reuters regarding the matter. The surge in the adoption of generative artificial intelligence technology has resulted in a significant increase in demand for AI data centers. These data centers need to be equipped with the capability to manage more complex tasks compared to traditional data centers.According to The Information’s report, Microsoft is anticipated to finance the project, which is projected to be around 100 times more expensive than some of the largest existing data centers. This information was gleaned from individuals familiar with private discussions about the proposal. The proposed supercomputer, based in the United States, would be the largest in a series of supercomputers that the companies intend to construct over the next six years.

The Information attributed the estimated cost of US$100 billion to individuals who spoke with Altman about it and to someone who has seen some of Microsoft’s preliminary cost estimates, although it did not disclose the identities of these sources. Altman and Microsoft have organized the supercomputers into five phases, with “Stargate” designated as the fifth phase. Additionally, Microsoft is collaborating on a smaller, fourth-phase supercomputer for OpenAI, slated for launch around 2026, as per the report.

Microsoft and OpenAI are currently progressing through the third phase of their five-phase plan, with a substantial portion of the expenses for the subsequent two phases earmarked for acquiring the necessary AI chips, according to the report. AI chips are typically sold at premium prices, with Nvidia CEO Jensen Huang indicating to CNBC in March that the latest “Blackwell” B200 artificial intelligence chip is expected to be priced between US$30,000 and US$40,000.In addition, Microsoft unveiled a pair of custom-designed computing chips in November of last year. The new project, as outlined in the report, is intended to be compatible with chips from various suppliers, reflecting a strategic approach to hardware integration and flexibility.

In their statement to Reuters, the Microsoft spokesperson highlighted the company’s continuous focus on planning for the next generation of infrastructure innovations aimed at advancing AI capabilities. This commitment underscores Microsoft’s proactive approach to staying at the forefront of technological advancements in the field of artificial intelligence. While the spokesperson refrained from directly commenting on the reported launch of the Stargate supercomputer, their acknowledgment of ongoing efforts in infrastructure development signals Microsoft’s strategic investment in driving forward the boundaries of AI research and application.

The reported expenses for the plan could surpass US$115 billion, which is more than triple the capital spending Microsoft allocated last year for servers, buildings, and other equipment. This significant investment underscores the scale and ambition of Microsoft and OpenAI’s collaboration in advancing artificial intelligence capabilities through the development of cutting-edge infrastructure.