The IPO numbers are done. SpaceX raised $75 billion on Nasdaq on June 12, surpassing Saudi Aramco’s 2019 record and pricing at $135 per share. Shares reportedly opened at $150 and closed at $160.95 according to market reports, pushing the company’s market capitalization to approximately $2.1 trillion. Three briefs in this hub covered that story as it happened. This one covers what the money is for.
The use-of-proceeds section of SpaceX’s S-1 filing with the SEC describes “deployment of in-orbit AI compute infrastructure” as a direct application of IPO proceeds. That’s not a roadmap slide. It’s a disclosure document. The difference matters: S-1 language is reviewed by underwriters and the SEC. SpaceX committed, in writing, to building AI compute infrastructure in orbit.
Secondary reporting adds context the S-1 excerpt doesn’t fully confirm, specifically, a solar-powered architecture for the orbital data centers. The full S-1 text should be reviewed before treating that framing as settled. What is settled, per the filing: SpaceX is building in-orbit AI compute, and the $75B is partly how it gets funded.
The real story is what that compute layer actually means structurally. Terrestrial AI data centers face three constraints: power grid availability, physical land permitting, and jurisdiction. Orbital infrastructure sidesteps all three. It also introduces its own constraints – latency, deployment cost, regulatory oversight from the FCC and ITU, but the energy and permitting problems that have slowed terrestrial data center builds don’t apply the same way.
BlackRock reportedly placed an anchor order of at least $5 billion, according to Bloomberg reporting cited by financial media. That’s a signal about institutional conviction, not just SpaceX’s story. When the world’s largest asset manager anchors an IPO at that scale, it’s telling other allocators something about how it’s categorizing this asset: not aerospace, not defense, not tourism. Infrastructure.
This is the third major AI infrastructure capital event in roughly 30 days. Oracle raised $40 billion. Apollo and Blackstone committed $35 billion in private credit toward AI infrastructure. The UK government announced a £6 billion package. SpaceX’s $75B is larger than all three combined. And it’s the only one pointing infrastructure investment off the ground.
Watch the FCC and ITU licensing timeline. SpaceX’s in-orbit compute thesis depends on spectrum allocation and satellite operation approvals that don’t follow the same clock as a data center build permit. Sovereign blockage of Starlink connectivity in key markets is the scenario that most directly undermines the infrastructure thesis, that risk isn’t theoretical, and it isn’t in the S-1 in reassuring terms.
The question IPO investors largely haven’t asked yet: what happens to the AI infrastructure thesis if the orbital layer faces regulatory fragmentation by jurisdiction? The terrestrial data center model has a 40-year playbook for that problem. In-orbit compute doesn’t. Watch for analyst coverage that engages this question specifically, it’ll separate investors who bought the record from investors who read the filing.