Microsoft and OpenAI's Stargate: The $100 Billion AI Supercomputer Venture

In a groundbreaking collaboration that could redefine the landscape of artificial intelligence, Microsoft and OpenAI have set the tech world abuzz with reports of a joint venture to create a supercomputer named ‘Stargate.’ With an estimated cost of up to $100 billion, this project is not just a technological leap but a financial titan in the making.

“Dota2 OpenAI戰隊打敗人類原因曝光 AI還是靠「作弊」取勝” by steamXO is licensed under CC PDM 1.0

The initiative, which is part of a five-phase plan, is expected to culminate with Stargate becoming operational by 2028. This U.S.-based supercomputer is just one of several installations planned over the next six years, showcasing the rapid adoption of generative AI technology and the increasing demand for data centers capable of handling advanced AI tasks.

While neither Microsoft nor OpenAI have officially commented on the project, sources close to the matter suggest that Microsoft will likely finance the venture. This isn’t surprising, given Microsoft’s previous commitment of over $13 billion to OpenAI and the exclusive rights it holds to resell OpenAI’s technology. The partnership has been so integral that some insiders worry Microsoft is becoming an IT department for the startup.

The sheer scale of Stargate is staggering. It’s projected to be 100 times more expensive than the largest data centers currently in operation, and it could require several gigawatts of power—so much so that alternative power sources like nuclear energy are being considered. The quest for server chips necessary for the AI upgrade is a primary factor driving the cost, with the demand for AI chips reaching a fever pitch. OpenAI CEO Sam Altman has even expressed interest in creating his own chips to mitigate reliance on the limited market, primarily controlled by Nvidia.

The technical challenges are as immense as the financial ones. Microsoft is tasked with figuring out how to pack more GPUs into a single rack to boost performance without causing overheating. The location of Stargate is still undecided, with possibilities ranging from a single data center to multiple centers in close proximity.

Microsoft’s involvement hinges on OpenAI’s promise to enhance its AI capabilities. With the current focus on procuring AI chips, the project is already in its third phase. The fourth phase involves a smaller supercomputer set to launch around 2026, with Stargate as the grand finale.

The implications of Stargate are vast. It has the potential to far exceed the computing power currently supplied by Microsoft to OpenAI, setting a new standard for AI infrastructure. As the companies navigate the complexities of chip acquisition and power sourcing, the industry watches with bated breath.

‘We are always planning for the next generation of infrastructure innovations needed to continue pushing the frontier of AI capability,’ a Microsoft spokesperson stated, encapsulating the forward-thinking ethos driving this venture. With the potential expenses exceeding $115 billion, more than triple Microsoft’s last year’s capital spending, Stargate represents not just a technological marvel but a bold bet on the future of AI.