SpaceX has more neocloud revenue
SpaceX's AI revenue grew more than three times to $2.6 billion from the year before, mostly because of deals that the company made to provide compute to other AI companies, according to SpaceX's quarterly earnings. The AI division, which the company said in its documents to go public was the source of most of its value, lost $1.5 billion this quarter, slightly less than in the same quarter last year.
SpaceX is now a neocloud with rockets. That is the cleanest way to read the latest earnings disclosure, in which AI-related revenue tripled to $2.6 billion while the division still bled $1.5 billion for the quarter. The growth came from compute deals with Anthropic and Google, placing the company in direct competition with CoreWeave and the rest of the rented-GPU economy.
The economics are familiar. Neoclouds buy capacity, package it, and resell it to model builders who need inference and training scale faster than they can build it themselves. SpaceX has a structural advantage the others do not: its own constellation of satellites and a launch cadence that lets it deploy and replace compute hardware on its own schedule. Vertical integration is the moat, and the launch manifest is the supply chain.
The loss is the tell. A $1.5 billion quarterly deficit on $2.6 billion in revenue implies heavy capital expenditure flowing through the income statement, the classic signature of a buildout phase. SpaceX is pricing compute aggressively to lock in anchor tenants before competitors can match the latency or the orbital data routing story. The bet is that owning the pipe and the silicon underneath it will eventually compress costs enough to flip the division profitable.
For the broader AI infrastructure market, the signal is that hyperscale buyers are diversifying away from a single provider. Anthropic and Google both signing with SpaceX means the neocloud tier is no longer a CoreWeave monoculture. That should keep pricing pressure on the entire rented-compute stack and slow the easy-money phase for anyone whose only asset is a warehouse full of GPUs.
The remote-work angle here is thin. SpaceX is hiring engineers and operations staff for its AI division, but the story is about compute supply, not labor markets. Worth noting only because the buildout will pull talent into a handful of U.S. data center hubs, tightening an already competitive AI engineering market.