Microsoft and OpenAI have been "drawing up plans" for a data center project to feature an AI supercomputer that is codenamed "Stargate" with millions of next-gen, specialized server chips to power OpenAI's artificial intelligence.
The news comes from The Information and "three people who have been involved in the private conversations about the proposal." According to a person who spoke to OpenAI founder and CEO Sam Altman about it and had viewed some of Microsoft's initial cost estimates, the new data center and AI supercomputer codenamed "Stargate" would cost as much as $100 billion to build.
The gigantic explosion of AI across virtually every industry has seen demand for AI data centers capable of handling massive use compared to traditional data centers is seeing multiple big players announcing and building new AI-focused data centers.
- Read more: Amazon to spend $150 billion on datacenters for 'explosion in demand' for AI
- Read more: Amazon announces Q just days after ChatGPT maker OpenAI announces Q*
Amazon recently penciled down $150 billion across the next 15 years for data centers, as they expected an "explosion in demand" for all things AI in the future.
Microsoft is expected to foot the $100 billion bill for the new "Stargate" AI supercomputer, which means it will be 100x more expensive than some of the biggest, most powerful data centers on the planet. The US-based AI supercomputer would be the biggest on the market and would be built by 2030.
- Read more: OpenAI Sora video tool uses 720,000 NVIDIA H100 GPUs worth $21.6 billion
- Read more: OpenAI is building a global network of dedicated chip fabrication plants
- Read more: Amazon AWS will join Google and Microsoft with data centers in Taiwan
- Read more: Microsoft announces chip design based on the new Intel 18A process node
Microsoft and OpenAI are expected to build the supercomputers in five phases, with the fifth phase being the final one with Stargate. According to The Information, the companies are reportedly working on a smaller, fourth-phase supercomputer for OpenAI that will be announced in 2026.
The two companies are reportedly in the middle of the third phase right now, with a "significant portion" of the cost of the next two phases leading into Stargate will be scooping up all of the required AI chips to build the AI supercomputers.
NVIDIA just announced its next-generation Blackwell B200 AI GPU. CEO Jensen Huang told CNBC that it would cost between $30,000 and $40,000 per AI GPU. Microsoft recently announced two new custom-designed chips in late 2023 and a new chip design based on Intel's upcoming Intel 18A process node.
A Microsoft spokesperson said in an email response to Reuters: "We are always planning for the next generation of infrastructure innovations needed to continue pushing the frontier of AI capability".
The Information reports that the AI supercomputer project towards its fifth phase codenamed "Stargate" could exceed $115 billion, more than 3x the expenditure in 2023 that Microsoft spent on servers, buildings, and other equipment.