A massive new AI infrastructure project dubbed “Stargate” just got announced in dramatic fashion at the White House—and it’s turning heads for a few reasons.
OpenAI, SoftBank, and Oracle are teaming up on this venture, which aims to invest up to $500 billion over four years to build colossal AI data centers across the US. The initial deployment alone comes in at a tidy $100 billion. According to the announcement:
This thing is so big that its backers say it could eventually surpass the inflation-adjusted cost of NASA’s Apollo program—and it’s got the potential to completely reshape the AI landscape. But, there's a lot of drama behind the scenes.
Beyond the headlines and hype, is Stargate for real?
I got all the details from Marketing AI Institute founder and CEO Paul Roetzer on Episode 132 of The Artificial Intelligence Show.
Compute power is the beating heart of frontier AI research. The largest AI models require jaw-dropping amounts of energy, money, and chips.
That’s where Stargate comes in. This new entity—reportedly a separate venture with OpenAI holding a large equity stake—plans to build multiple massive data centers dedicated exclusively to OpenAI’s needs. That means:
If it all comes together, Stargate would dwarf most existing AI infrastructure projects in scale—and mark a major shift in how OpenAI (and possibly other AI companies) source the computing muscle they need.
Here’s where it gets messy. That enormous $500 billion price tag might be more of a long-term target than a locked-in budget. OpenAI doesn’t actually have $500 billion lying around, so there’s plenty of speculation on how (and whether) they’ll raise the full amount.
Industry insiders are also skeptical about supply issues:
“It’s very apparent OpenAI doesn’t actually have this kind of money,” says Roetzer. “They’re going to have to raise it on their balance sheet.”
As if the financial hurdles weren’t complex enough, Stargate also reveals new tensions between OpenAI and its longtime deep-pocketed partner, Microsoft.
Then Satya Nadella waded into the conversation, joking that his own $80 billion is real, implying others might be playing fast and loose with the numbers. Elon Musk replied in agreement, firing a shot across Altman’s bow about hyping “unreal” AI investments.
It’s part soap opera, part trillion-dollar business dance—an ongoing saga that underscores how big AI has become (and how many big egos are now in play).
Stargate got unveiled just a day after President Trump’s inauguration, signaling a push for massive AI infrastructure growth in a deregulated environment. That lines up with what Roetzer has been predicting for some time:
“We said on the podcast many times it’s all about deregulation and acceleration of technology. How it plays out, who the winners are, that’s the question.”
Trump is no fan of wind or solar, so it’s unclear how that might influence the energy sources powering Stargate’s data centers. Meanwhile, other industry leaders like Elon Musk remain champions of renewables. The result: a storm of political, financial, and technical factors all colliding at once.
No matter how this shakes out though, there's one trend you can bet on, says Roetzer:
“Infrastructure and energy is going to drive everything in America for this coming half decade, if not more.”
Though the final scale and execution of Stargate remains up in the air, the broader implications are clear:
Despite its issues, Stargate does show a willingness by some AI leaders to commit to large-scale, moonshot-style funding for the infrastructure side of the AI equation. In that sense, it’s a major milestone—one that has the potential to shape the future of AI for a generation.
But given the swirling questions—about who’s truly paying, whether resources exist to execute, and how this affects relationships across the industry—nobody’s popping champagne just yet.