Starcloud Raises $170M Series A to Build Data Centers in Orbit

Every so often, tech erases its own ceiling without asking for permission. Starcloud just did that, pulling in $170M in a Series A and touching a $1.1B valuation. Earth is starting to look like a crowded place for compute, and Starcloud is already working above the noise.

Philip Johnston, Ezra Feilden, and Adi Oltean didn’t just launch a company in January 2024, they introduced a new coordinate system for infrastructure. Redmond, Washington is the mailing address, but the real operation lives in low Earth orbit, where power flows uninterrupted, cooling comes naturally, and the usual constraints don’t get a vote.

Benchmark and EQT Ventures co-led the round, with Macquarie Capital, NFX, Nebular, Y Combinator, Adjacent, 776, Fuse, Manhattan West, and Monolith Power Systems stepping in with conviction. When a syndicate like that aligns, it usually signals more than optimism. Chetan Puttagunta joining the board adds another layer of intent. This is infrastructure being built with expectation, not curiosity.

Starcloud is building orbital data centers, and they are already in motion. Starcloud-1 launched in November 2025, carrying an NVIDIA H100 GPU into orbit and running real workloads, including training a language model in space. Not theoretical. Not someday. Already executed.

The premise cuts clean. Terrestrial data centers are straining under power demand, land constraints, and cooling complexity. Orbit offers continuous solar energy and a vacuum that handles heat without negotiation. Starcloud isn’t forcing innovation through resistance, they are placing it where the physics cooperate.

The $170M accelerates that direction. Starcloud-2 is lined up for 2026 with a massive leap in power and the largest commercial deployable radiator sent to space. Crusoe is preparing to run workloads, with AWS, Google Cloud, and NVIDIA in the orbit of this ecosystem. That is early infrastructure being claimed before the rest of the market fully processes what’s happening.

The real signal sits in execution. Move quickly, prove capability, and collapse the distance between concept and reality. Starcloud didn’t wait for consensus, they delivered results from orbit and let the market catch up afterward. Congratulations to Philip Johnston, Ezra Feilden, Adi Oltean, and the entire Starcloud team. Scaling compute used to be about stacking servers. Now it is about changing altitude, literally and strategically, at the same time.