- LEO (Low Earth Orbit)
-
Low Earth Orbit refers to the orbital region between approximately 160 to 2,000 kilometers (100 to 1,240 miles) above Earth's surface, where satellites complete one orbit every 90-120 minutes. This altitude range offers significant advantages for data center applications: lower launch costs compared to higher orbits, reduced communication latency (approximately 20 milliseconds round-trip), and sufficient atmospheric drag to enable passive deorbiting for space debris mitigation. LEO sits below the Van Allen radiation belts, reducing but not eliminating radiation exposure. Most earth observation satellites, the International Space Station (orbiting at ~400km), and planned orbital data center constellations operate in LEO due to these combined benefits of accessibility, performance, and cost-effectiveness.
- ODC (Orbital Data Center)
-
Orbital Data Center (ODC) is the technical term for computing facilities deployed on satellites or space platforms in Earth orbit, performing data processing, storage, and transmission functions traditionally handled by terrestrial data centers. ODCs leverage space environment advantages including continuous solar power, passive radiative cooling, and proximity to satellite-generated data sources. The term encompasses various architectures from dedicated computing satellites (like Starcloud's constellation approach) to integrated modules on crewed space stations (like Axiom Space's pressurized ODC nodes). ODCs are distinguished from traditional satellites by their primary mission of general-purpose computing services rather than specific functions like communications relay or earth observation, though many serve hybrid roles processing data from co-located sensors.
- Space-Based Computing
-
Space-based computing represents the broader category of computational infrastructure deployed beyond Earth's atmosphere, including orbital data centers, satellite edge processors, lunar data facilities, and deep-space computing systems. This encompasses everything from simple microcontrollers managing satellite functions to sophisticated AI training clusters and cloud services accessible from Earth. The term highlights the paradigm shift from treating space systems as remote clients of terrestrial computing (uploading commands, downloading data) to deploying substantial processing capabilities in space itself. Applications span on-orbit satellite data processing, autonomous spacecraft operations, scientific research computing on space stations, and increasingly, commercial cloud services leveraging space's unique physics for energy efficiency and global coverage.
- Edge Computing
-
Edge computing is the distributed computing paradigm that processes data near its source rather than centralizing computation in distant data centers, reducing latency, bandwidth consumption, and enabling real-time decision-making. In space contexts, edge computing becomes critical: satellites generating terabytes of imagery cannot transmit all raw data to Earth, requiring on-orbit processing to extract actionable intelligence before downlink. LEO data centers represent the ultimate edge architecture—co-locating compute resources with data-generating sensors, eliminating hours-long ground station wait times, and reducing bandwidth requirements by 60-90%. Space edge computing also supports autonomous operations for deep-space missions where communication delays (minutes to hours) make Earth-based control impractical, requiring spacecraft to process sensor data and make decisions locally.
- Satellite Constellation
-
A satellite constellation is a coordinated group of multiple satellites working together to provide continuous global or regional coverage, overcoming the limited field-of-view and intermittent ground contact windows of individual spacecraft. Orbital data center constellations like Starcloud's planned 300-satellite network employ this architecture to ensure any point on Earth can access computing services at any time. Satellites in the constellation communicate via optical inter-satellite links, forming a mesh network that routes data between spacecraft to reach those currently over ground stations. Constellation design involves complex orbital mechanics: planes at different inclinations, phasing satellites within planes for coverage gaps, and altitude selection balancing atmospheric drag, radiation exposure, and communication latency. This approach enables LEO systems to match geostationary satellite's continuous availability while maintaining LEO's latency and launch cost advantages.
- Ground Station Network
-
Ground station networks are geographically distributed facilities with antennas and optical terminals that communicate with satellites, providing the interface between space-based infrastructure and terrestrial internet connectivity. For orbital data centers, ground stations serve three critical functions: uploading user workloads and commands, downloading processed results, and providing network connectivity for cloud services accessed from Earth. Optical ground stations require clear weather and line-of-sight, necessitating multiple sites at high-altitude locations (mountaintops, deserts) for redundancy. Radio frequency ground stations operate through clouds but offer lower bandwidth (100-500 Mbps versus 10+ Gbps for optical). The ASCEND project specifies "ground nodes interfacing with the internet," acknowledging that extensive ground infrastructure remains necessary despite space-based processing. Commercial providers like AWS Ground Station and Azure Orbital offer ground station networks as a service, reducing infrastructure barriers for satellite operators.
- Radiation Hardening
-
Radiation hardening encompasses design techniques and technologies that protect electronic systems from the damaging effects of cosmic radiation, solar particle events, and trapped radiation belts in space. Traditional approaches use specialized semiconductor fabrication processes creating radiation-hardened chips resistant to single-event upsets and total ionizing dose effects, but these components cost 600x more than commercial equivalents and lag a decade behind state-of-the-art performance. Modern software-based hardening, demonstrated by HPE's Spaceborne Computer, employs error-detecting memory, triple modular redundancy (three processors vote on results), and adaptive throttling during solar storms, enabling commercial off-the-shelf components to survive years in orbit. Physical shielding using aluminum, tungsten, or specialized composites provides additional protection but adds launch mass. For LEO data centers to use cutting-edge AI accelerators like NVIDIA H100 GPUs, software hardening and strategic component placement prove essential for balancing performance, cost, and reliability.
- Thermal Management (Space)
-
Thermal management in space relies exclusively on radiative heat transfer—emitting infrared radiation from deployable radiator panels to the 2.7 Kelvin vacuum—since convection cooling using air or water circulation is impossible without atmosphere. Heat generated by processors transfers via liquid cooling loops (typically ammonia or water-glycol) to large black-surfaced radiator plates positioned to face deep space rather than the sun. Cooling capacity scales with radiator surface area and operating temperature to the fourth power per Stefan-Boltzmann law: a 100-kilowatt data center requires approximately one basketball court of radiators. Challenges include extreme temperature cycling (150+ degree swings between sunlight and shadow every 90-minute orbit), managing heat in vacuum where conduction and convection don't work, and deploying square-kilometer radiator arrays for gigawatt facilities. The ISS uses 477 square meters of ammonia-filled radiators to dissipate 75 kilowatts, providing heritage for scaling orbital data center cooling systems.
- Launch Economics
-
Launch economics determine orbital data center viability through the cost per kilogram to reach orbit, which has declined 95%+ from Space Shuttle-era $54,500/kg to current SpaceX Falcon 9 prices around $1,500-$2,720/kg through reusability and commercial competition. Further reductions are projected: SpaceX Starship targets $100/kg through full reusability and 100+ tonne payload capacity, while CitiGPS forecasts industry costs reaching $33/kg by 2040. These improvements transform business cases: deploying a 40-megawatt orbital cluster costs $8.2 million over 10 years at $30/kg launch pricing (Starcloud projection) versus $167 million for terrestrial equivalents dominated by electricity expenses. However, gigawatt-scale ASCEND facilities require launching 5,000-8,000 tonnes—feasible at $100/kg ($500-800M launch costs) but economically prohibitive at current prices ($7.5-21.7B). Launch cost trajectory represents the primary variable determining whether orbital data centers remain niche applications or achieve mainstream adoption.
- Data Sovereignty (Space)
-
Data sovereignty in space contexts addresses jurisdictional questions about where data is legally stored and processed when facilities orbit beyond any nation's territory, creating unique regulatory and security implications. The European ASCEND project explicitly targets data sovereignty as a core objective—enabling the EU to maintain independent computing capability without reliance on US or Chinese cloud providers whose terrestrial data centers are subject to foreign legal frameworks like the CLOUD Act. Orbital facilities in international space aren't subject to terrestrial search warrants or data access requirements, though launch nation regulations and ITU/UN space treaties apply. Questions remain unsettled: What jurisdiction governs data breaches on orbital platforms? Can governments compel access to data stored in space? How do export controls apply to hardware with dual-use applications? As one analysis notes: "International laws and regulations governing tech in space are still evolving," creating both opportunities for sovereign control and regulatory uncertainties requiring legal frameworks to mature alongside the technology.
- Optical/Laser Communication
-
Optical communication systems use laser beams to transmit data between satellites (inter-satellite links) and between space and ground (downlinks/uplinks), offering 10-100x higher bandwidth than radio frequency alternatives while providing inherent security through narrow beam propagation difficult to intercept. Current systems achieve 2.5-10+ Gbps per link, with NASA's TBIRD demonstration proving 200 Gbps is feasible. Axiom Space's Orbital Data Center nodes integrate with Kepler Communications' optical relay network providing 2.5 Gbps links compatible with Space Development Agency standards, with roadmaps targeting terabit-per-second aggregate throughput. Advantages include immunity to radio frequency interference, no spectrum licensing requirements, and higher power efficiency than RF. Challenges involve precise pointing requirements (laser beams spread to only meters across after hundreds of kilometers), atmospheric interference affecting ground links during clouds or turbulence, and acquisition/tracking complexity. Optical inter-satellite links enable mesh networking architectures where satellites relay data through neighbors, providing continuous connectivity impossible with ground-station-dependent RF systems.
- Space Debris
-
Space debris encompasses defunct satellites, spent rocket stages, collision fragments, and paint flecks traveling at orbital velocities (7-8 km/second in LEO), posing collision risks to operational spacecraft and potentially triggering cascading Kessler Syndrome where debris-generating impacts render orbits unusable. The growing debris population—tracked objects exceed 34,000 larger than 10cm with millions of smaller untracked particles—creates significant risks for orbital data centers requiring long operational lifetimes. Mitigation strategies include designing satellites for post-mission disposal (deorbiting within 25 years), active debris removal missions, collision avoidance maneuvers guided by space surveillance networks, and shielding critical components. LEO orbits below 600km benefit from atmospheric drag that naturally deorbits debris within decades, while higher altitudes require propulsive deorbit capabilities. The ASCEND project identified space debris as a key risk factor requiring "comprehensive debris mitigation strategies" as constellation sizes grow into hundreds or thousands of satellites.
- GEO (Geostationary Earth Orbit)
-
Geostationary Earth Orbit refers to a circular orbit approximately 35,786 kilometers above Earth's equator where satellites complete one orbit in exactly 24 hours, matching Earth's rotation and appearing stationary above a fixed point. While GEO provides continuous coverage of nearly half the planet from a single satellite—ideal for communications and weather monitoring—it offers poor characteristics for data centers compared to LEO: significantly higher launch costs, 500+ millisecond round-trip latency unsuitable for real-time applications, and exposure to intense Van Allen radiation belts requiring expensive radiation hardening. GEO satellites also face challenges with propellant-limited stationkeeping and permanent debris accumulation in crowded orbital slots. Current orbital data center initiatives focus on LEO for its lower costs, reduced latency, and superior economics.
- MEO (Medium Earth Orbit)
-
Medium Earth Orbit encompasses altitudes between approximately 2,000 and 35,786 kilometers, positioned between Low Earth Orbit and Geostationary Orbit. GPS, Galileo, and GLONASS navigation satellite constellations operate in MEO (typically 20,000-23,000km), balancing broader coverage footprints than LEO against lower latency and launch costs than GEO. For data center applications, MEO offers few advantages over LEO—higher radiation exposure, increased launch costs, and longer communication latency—while lacking GEO's continuous single-satellite coverage. The orbital data center industry has largely bypassed MEO in favor of LEO constellations that provide global coverage through multiple satellites while maintaining low latency and lower deployment costs.
- PUE (Power Usage Effectiveness)
-
Power Usage Effectiveness is the industry-standard metric for data center energy efficiency, calculated as total facility power consumption divided by IT equipment power consumption. A PUE of 1.0 represents perfect efficiency where all energy powers computing hardware, while typical terrestrial data centers achieve 1.3-1.6 PUE with substantial energy lost to cooling systems, power conversion, and lighting. Orbital data centers could theoretically approach 1.0 PUE by eliminating mechanical cooling (using passive thermal radiation), operating in constant moderate temperatures, and accessing solar power directly without grid losses. However, realistic PUE accounting must include substantial power for attitude control, communications, thermal management pumps, and electronics for autonomous operations. The true efficiency advantage comes not from PUE improvements but from accessing abundant zero-carbon solar energy unavailable to terrestrial facilities constrained by grid capacity and land availability.