Microsoft and OpenAI are reportedly joining forces to create a state-of-the-art AI supercomputer in the United States, known by its codename, Stargate. Sources close to the matter have revealed that this supercomputer is expected to be the most significant among several facilities planned for the next six years.
The Stargate project aims to host millions of specialized chips, designed to enhance the AI capabilities of OpenAI significantly. Microsoft, a key investor in OpenAI, with an investment exceeding $13 billion, is set to finance this grand venture.
The collaboration between Microsoft and OpenAI isn’t new. They’ve been in the tech tango for a while, working on a smaller-scale supercomputer in Mt. Pleasant, Wisconsin. This earlier project, with a budget touching $10 billion, is speculated to be powered by Nvidia’s AI chips.
Yet, Stargate is where the vision expands, with plans to integrate GPUs from various manufacturers, including AMD and Microsoft’s own server chips. This strategy aims to lessen reliance on a single supplier, diversifying the tech giant’s technological arsenal.
The financial roadmap for Stargate is staggering. Before reaching the final phase of this initiative, the investment is expected to dwarf the earlier $100 billion projection, pushing the total expenditure beyond $115 billion. To put this in perspective, this figure is triple what Microsoft spent last year on servers and other infrastructure.
The project’s launch is aimed for 2028, with full operational capacity expected by 2030. Given the supercomputer’s vast energy requirements, estimated to be several gigawatts, the planning team is considering alternative energy sources, including nuclear power, to meet its demands.
