AI compute is scarce on Earth
Demand for inference and training is rising exponentially, while permitting and grid connections delay new data centers by years. Orbit offers capacity on a faster timeline.
Space Data Centers
Bringing intelligence to orbit
Space-graded AI compute, launched and operated at scale, for the workloads Earth can no longer wait on.
Backed byBuilding withBacked by ·
Building with
Building with
DIHP
Grant award · 2026
Fraunhofer IOF
Hardware prototyping · Jena
Entrepreneurs First
Deep tech · 2026 cohort
Why it matters
The bottleneck is no longer launch or sensing. It's compute on Earth and in orbit.
Sovereign compute isn't on demand
Governments need to run sensitive AI workloads on infrastructure that stays out of reach of adversaries: secure, dedicated, and available when it's needed.
Space data is stuck behind downlink
Constellations collect terabytes a day but can only downlink a fraction. Processing in orbit turns raw pixels into insights before they ever leave space.
Who it's for
AI labs & hyperscalers
MW-class capacity online years sooner than terrestrial buildout: compute-as-a-service from orbit.
Government & defence
Sovereign, secure compute placed physically out of reach for the sensitive workloads proliferating across national systems.
Earth observation
In-orbit processing that turns collected data into sellable insights before downlink: more value per pass.
Team
Tanmay Gupta
CEO · Co-Founder
Space instrumentation, hardware & photonics.Ex-NASA JPL, Caltech, ZEISS, Planet.
Vidhi Desai
CTO · Co-Founder
Spacecraft systems & machine learning.Ex-Viasat, Cranfield Rocketry.