March 31, 2024 – Microsoft and OpenAI executives are in discussions regarding an unprecedented data center project, according to foreign media reports. Dubbed “Stargate,” the ambitious initiative aims to construct an AI supercomputer powered by millions of specialized server chips. This colossal machine is intended to bolster OpenAI’s artificial intelligence technologies.
Insider sources have revealed that the investment scale for this undertaking could potentially reach an astonishing 100billion,equivalenttoapproximately722.13billionChineseyuan.Whenconsideringthebroaderscheme,thetotalinvestmentmightexceed115 billion, a figure that dwarfs Microsoft’s capital expenditure on servers, office buildings, and other equipment last year by more than threefold.
Microsoft’s willingness to proceed with the “Stargate” project hinges significantly on OpenAI’s ability to significantly enhance its AI capabilities. If the “Stargate” initiative progresses smoothly, it is expected to deliver an exponential surge in computing power.
However, the colossal undertaking faces a considerable challenge in securing adequate power supply. It is anticipated that “Stargate” will require a minimum of several billion watts of electricity, a figure comparable to the combined power consumption of multiple large-scale data centers currently in operation. As such, obtaining a sufficient power supply emerges as a significant hurdle.
Microsoft and OpenAI executives have discussed a potential launch date for “Stargate,” with the earliest being 2028 and the latest being 2030. By the time of its completion, the project might demand up to 5 gigawatts of electricity.
