Elon Musk, the world's first trillionaire after SpaceX's $85.7 billion IPO, is pitching orbital AI data centers as inevitable — and he needs you to believe it, because his company's trillion-dollar valuation depends on it, while the price tag for just the ground-side factory could hit $119 billion.
The stake for ordinary Americans is simple: while you're told there's no money for infrastructure or housing, the world's richest man is mobilizing capital and regulatory approvals for a million-satellite constellation in low-Earth orbit. The money flowing into this orbital fantasy isn't disappearing into space — it's being extracted from the real economy.
Musk claimed at an Austin event in March that space-based, solar-powered data centers will be more cost-effective than terrestrial ones in as little as two to three years. "Increasing power on Earth becomes harder over time and more expensive over time," he said, "but in space it becomes actually cheaper and easier over time." In January, SpaceX filed an FCC application for a constellation of up to one million so-called AI1 satellites — upgraded Starlink units requiring exponentially more semiconductors.
The manufacturing scale is staggering. SpaceX, Tesla, and Intel have partnered to build Terafab, a 10-million-square-foot facility in Austin that could cost up to $119 billion and isn't slated to open until 2029. That's more than the annual GDP of several U.S. states, committed to feed an orbital compute network that, by one investor's admission, doesn't yet make economic sense. Duncan Davidson, a partner at Bullpen Capital, told CNBC the engineering issues are being solved but conceded: "economically, right now, it's marginal." He added the business case is "really strong" — but only if Starship becomes operational and lowers launch costs, and if terrestrial data center costs keep climbing. That's a lot of ifs.
CNBC framed the story as a plausible next step driven by SpaceX's IPO windfall and the constraints facing Earth-bound data centers. Barchart buried the lede: Musk has a documented history of overpromising and underdelivering that should make investors "highly skeptical." He predicted coast-to-coast autonomous Tesla travel by 2017 — never happened. He promised one million robotaxis by 2020 — as of May 2026, exactly 39 unsupervised Tesla robotaxis were operating. The hyperloop was scrapped. Neuralink mass production was promised for 2026 with no delivery in sight.
Musk now claims SpaceX could generate $1 trillion in revenue by 2030. Barchart's assessment is blunt: the only way that target is reachable is if space-based data centers take off and xAI becomes a dominant supplier. Both are long shots in a field where Anthropic, OpenAI, Amazon, and Alphabet are already competing aggressively.
Seeking Alpha, meanwhile, ran a piece arguing that tower REITs like American Tower aren't threatened by satellites — conveniently disclosed that the author holds a long position in AMT. The real insight buried in that financial newsletter: satellites solve coverage, not capacity. The future, even by Wall Street's own analysis, is an integrated ecosystem where towers, fiber, and data centers still do the heavy lifting. The orbital pitch is, at best, a supplement — not a replacement.
Jeff Bezos and Alphabet are also in the race, per CNBC, which means this isn't just a Musk fantasy — it's a billionaire class project. The question isn't whether orbital data centers can work eventually. It's whether the public should subsidize, regulate, and finance a speculative infrastructure that serves the AI industry's insatiable compute demands while the ground beneath working Americans keeps eroding.
SpaceX declined to elaborate on its plans. The FCC is reviewing a million-satellite application. And the same man who promised you a self-driving car in 2019 is now asking you to fund his data centers in the sky. The valuation demands it. The question is whether anyone will demand something in return.
