OpenAI drops $400B on massive US data center plan

Just a heads up, if you buy something through our links, we may get a small share of the sale. It’s one of the ways we keep the lights on here. Click here for more.

OpenAI is cranking its AI ambitions up to nuclear-reactor levels, literally. 

The ChatGPT maker just announced plans to pour a jaw-dropping $400 billion into five new US data center sites, a mega-build that could light up entire cities (7 gigawatts of power, to be exact) and bring it closer to its audacious half-trillion-dollar AI infrastructure goal. 

The sites, scattered across Texas, New Mexico, and Ohio, are part of Stargate, a sci-fi–sounding partnership with Oracle and SoftBank to keep OpenAI’s servers humming as ChatGPT balloons to 700 million weekly users.

The reveal came at a press event in Abilene, Texas, where OpenAI CEO Sam Altman looked every bit the AI evangelist. (Via: Bloomberg)

“We will push on infrastructure as hard as we can,” he promised, making it sound like a race to build the digital equivalent of a NASA space project. 

Three of the sites, including a secretive Midwest location, will team with Oracle to deliver more than 5.5 gigawatts of raw computing muscle, with an additional 600-megawatt expansion near Abilene. 

SoftBank will co-develop two more facilities, one in Lordstown, Ohio, and another in Milam County, Texas, marking its first hands-on Stargate project.

The price tag is staggering even by Silicon Valley standards. The $400 billion commitment follows a July deal with Oracle for $300 billion of Stargate capacity, and executives hinted that a fresh $100 billion pact with Nvidia will help finance the effort. 

“We haven’t figured out the final form of what financing for compute looks like,” Altman admitted, proving even AI’s biggest dreamers need spreadsheets.

This expansion isn’t just about bragging rights. Tech giants from Meta to Microsoft are already on track to spend more than $344 billion this year on AI data centers. 

With states and rivals racing to dominate the AI frontier, OpenAI wants enough horsepower to ensure no chatbot goes hungry. 

The company promises tens of thousands of US jobs along the way, and maybe a few nervous glances from anyone worried about an AI bubble. 

As Altman quipped, “We never want to be this compute-constrained again.”

Is OpenAI’s $400 billion data center investment a necessary step to maintain US leadership in AI, or does this massive spending suggest an unsustainable AI infrastructure bubble? Should we be concerned about the environmental impact of these energy-hungry facilities requiring 7 gigawatts of power, or do the potential benefits of advanced AI justify the resource consumption? Tell us below in the comments, or reach us via our Twitter or Facebook.