🚨 BREAKING: Microsoft and OpenAI plan supercomputer project worth $100 billion called ‘Stargate,‘ report says. The U.S.-based data center would reportedly house a supercomputer made up of millions of AI chips , according to a report by The Information on Friday.
The article stated a projected expense of $100 billion, referencing an individual who conversed with Sam Altman and another who reviewed preliminary cost projections. Additionally, there’s a strong indication that Microsoft would fund the endeavor, which could surpass the costs of some of the largest data projects by over a hundredfold and is anticipated to materialize by 2028.
According to the report, Stargate is slated to be the largest among a sequence of supercomputers that Microsoft and OpenAI intend to construct within the next six years. The supercomputer initiative is presently navigating through the third stage, with Stargate earmarked for the fifth and concluding phase. The report indicates that a substantial portion of the expenses for the forthcoming phases will be allocated towards procuring AI chips.
A bottleneck persists in the AI development pipeline due to the scarcity of graphics processing units (GPUs), essential for processing data in AI models. The surge in demand for these chips, driven by the generative AI explosion, has exacerbated the situation. Furthermore, Nvidia, a primary manufacturer renowned for its top-tier AI chips, holds nearly monopolistic control over the market. However, it faces challenges in meeting escalating demands, resulting in delayed orders. Concurrently, the cost of its chips continues to rise, adding to the complexity of the situation.
Following Nvidia’s GTC 2024 event, CEO Jensen Huang affirmed that the company’s newest AI chip, dubbed Blackwell B200, will be priced in the range of $30,000 to $40,000 per unit. With over $10 billion invested in its research and development, this chip significantly surpasses all competitors currently available on the market.
Presently, a consortium of technology firms, comprising Intel, Qualcomm, Google Cloud, Arm, and Samsung, has joined forces to establish the “United Acceleration Foundation.” This initiative introduces an open standard accelerator programming model, with the objective of disrupting Nvidia’s dominant position in both software and hardware realms within the AI domain.
Sampad