Microsoft and OpenAI are reportedly working on plans for a data centre project that would contain a supercomputer with millions of specialised server chips to power OpenAI’s artificial intelligence.
The project could cost as much as $100 billion, a report from The Information said.
Microsoft will be responsible for financing the project that executives envisage will bring to life the proposed U.S.-based supercomputer, which they refer to as Stargate.
The Stargate supercomputer will be the fifth phase of a series of supercomputers. The fourth-phase supercomputer for OpenAI is expected to be launched around 2026.
(For top technology news of the day, subscribe to our tech newsletter Today’s Cache)
The high cost of building the supercomputer is being attributed to bottlenecks in the supply of high performance GPUs used for the development of AI models.
Currently, Nvidia is one of the largest suppliers of high-end chips, and is struggling to keep up with ever-increasing demand.
Prices of AI chips have also increased with Nvidia’s latest release, the Blackwell B200 costing anywhere between $30,000-40,000.