This page may contain stale information. Last updated: 2026-06-12
Definition
Orbital data centers are artificial intelligence and general-purpose compute facilities deployed on satellites or space stations, designed to process workloads in orbit rather than transmitting raw data to terrestrial facilities.
SpaceX Roadmap
Per gwynne-shotwell and SEC filings (June 2026):
- Phase 1 (2026–2027): Add compute modules to existing Starlink broadband and Starlink Mobile satellites
- Phase 2 (late 2027): Launch dedicated AI1 satellites with purpose-built orbital AI infrastructure
- Phase 3: Scale to support 100,000+ satellite constellation with distributed inference
Net IPO proceeds allocate ~20% to AI/orbital data center infrastructure (2026-06-12-spacex-sec-s-1-amendment-debut).
Technical Challenges
- Radiation hardening: Space electronics require specialized shielding and fault tolerance
- Thermal management: Passive cooling in vacuum vs. active cooling on ground
- Latency: Orbital inference for edge workloads vs. ground round-trip delays
- Launch economics: Cost per kg to orbit must justify compute density
- Regulatory: Spectrum allocation, debris mitigation, export controls
Strategic Rationale
- Solar power availability in orbit (no grid dependency)
- Cooling via radiative heat dissipation in vacuum
- Proximity to Starlink constellation for low-latency edge AI
- Vertical integration with spacex launch and satellite manufacturing
SEC prospectus flags orbital AI data centers as "unproven technology with substantial technical, regulatory, and capital expenditure risks."