Starcloud raises $250M as it bets on orbital AI data centers

Starcloud has pulled in a $250 million extension to its Series A, pushing the orbital computing startup's valuation to $2.3 billion. The round adds to the $170 million the company raised in its original March 2026 Series A, and this time Nvidia put in $25 million directly.

Manhattan West Ventures led the extension. Cisco, Benchmark, EQT, Soma, NFX, 776, Cedar Capital, Goanna Capital, and Standard Capital also joined, a roster that spans traditional venture firms and strategic corporate investors betting on compute moving off the ground and into orbit.

Starcloud's pitch is straightforward: put AI inference hardware on satellites, where solar power is abundant and cooling doesn't require the water and land that terrestrial data centers consume. CEO Philip Johnston has positioned the company as an early mover in a category that barely existed two years ago.

The near-term plan involves two 8 kW compute satellites, called Starcloud-2, launching on rideshare missions in 2027 to run orbital inference workloads for US government agencies. That's a modest test of the concept compared to what comes next.

A bigger spacecraft, and a bet on Starship

Starcloud-3, described as a full orbital data center spacecraft, is designed to fly on SpaceX's Starship rocket rather than the smaller Falcon 9. That timing matters: SpaceX intends to phase out Falcon 9 by 2028 in favor of Starship, which promises far greater payload capacity to orbit.

Betting the company's flagship hardware on a rocket that hasn't yet proven itself at the cadence and reliability of Falcon 9 is a real dependency. If Starship's timeline slips, so does Starcloud-3.

The company has also filed with the FCC for permission to operate 88,000 spacecraft, a figure that dwarfs its current two-satellite roadmap and signals ambitions toward a full orbital compute constellation rather than a handful of demonstration units. Whether regulators, launch capacity, or customer demand catch up to that number is a separate question from whether the technology works.

For now, Starcloud's business rests on proving that inference workloads can run reliably in orbit, that government customers will pay for it, and that the launch vehicles it needs will show up on schedule. The funding gives it runway to try; it doesn't remove any of those variables.