Where Should We Put an Orbital Data Center?

Part 10 of 30 · Series date: September 8, 2026

The right address must suit the sunlight, the workload, the customer, and the route out at retirement.

Choosing an orbit resembles choosing a factory site, except the property keeps moving and cannot ignore celestial mechanics. A location that is excellent for sunlight may be awkward for customers. A location with convenient coverage may impose a long communications path.

There is no universally best address above Earth.

Close to home

Low Earth orbit keeps spacecraft relatively near the surface. That can help with transportation and communication distance. But a satellite does not remain over one customer. Coverage requires appropriate passes, relays, or multiple spacecraft.

Atmospheric drag, eclipse patterns, traffic, and disposal must also enter the design. Conditions vary greatly with altitude. Treating all low orbits as interchangeable hides important differences.

ESA's orbit guide distinguishes low, medium, polar, sun-synchronous, and geostationary arrangements and the purposes they serve. ESA: types of orbits.

Following the lighting schedule

A sun-synchronous orbit is arranged so that its plane maintains a particular relationship with the Sun as Earth travels around it. Such orbits are useful for repeatable observation lighting. Certain dawn-dusk designs can offer long sunlight exposure.

However, “sun-synchronous” does not mean “always illuminated.” Eclipse behavior depends on the complete orbit and season. Designers must calculate it, not infer it from the label. ESA: polar and sun-synchronous orbit.

For an imagined batch-computing service, a predictable energy schedule might be acceptable. For a service promising continuous operation, the same shadows could require storage, redundancy, or more limited customer commitments.

Appearing to stand still

A geostationary spacecraft circles above the equator at approximately 35,786 kilometers altitude, matching Earth's rotation so it appears fixed in the sky. This can simplify continuous visibility from suitable ground locations. It also places the spacecraft much farther away than a low-orbit node.

NASA's orbital mechanics introduction helps distinguish geosynchronous motion from the more restrictive geostationary case. A geosynchronous orbit matches Earth's rotation relative to the stars: about 23 hours and 56 minutes, a sidereal day. Only the circular, equatorial, prograde geostationary case stays approximately fixed in the sky for a ground observer. NASA: planetary orbits.

Greater distance affects launch requirements, path loss, delay, and servicing. Even geostationary platforms experience seasonal eclipse periods; a fixed appearance is not a guarantee of uninterrupted solar generation.

Radiation and customers complicate the map

The space environment is not uniform. Radiation exposure depends on trajectory, shielding, solar events, and operating duration. Neither “higher” nor “lower” is a complete radiation assessment.

Customers have geography too. A spacecraft processing nearby instrument data may benefit from matching the instrument's region of operation. A service for terrestrial users may favor different coverage. A lunar customer changes the problem again.

The best site is therefore a relationship among the power source, computers, data producers, and receivers—not a point chosen for one attractive feature.

The Moon is another category

A lunar surface facility has a physical foundation, but also dust, terrain, thermal cycles, construction difficulty, and substantial communication distance from Earth. Polar lighting requires local terrain analysis. A ridge and an adjacent crater can offer radically different conditions.

Placing equipment on the Moon should be justified by lunar work or a demonstrated special advantage. Merely calling it a stable platform does not establish that it is an economical Earth cloud region.

Three customers searching for an address

Imagine three design teams choosing an orbit. The first wants to process imagery from a particular observation mission. The second wants to run flexible computing jobs using long periods of sunlight. The third wants a power transmitter to remain visible from a fixed receiving region.

Their priorities differ before they choose a single component. The imagery team cares about where its inputs can arrive. The batch-computing team may accept waiting for ground delivery but needs a predictable energy and heat schedule. The power team must consider receiver visibility and beam geometry alongside its energy supply.

A location that wins for one team may lose badly for another. Combining their favorite features into one description—near Earth, almost always sunlit, fixed above a receiver—does not make those features physically compatible in one orbit.

This is why a proposal should provide an actual orbit or family of orbits rather than the broad label “space.” The address is part of the machine.

Altitude is not communications distance

A spacecraft's altitude measures height above a reference surface. Its path to a ground station is usually slanted, and a message may pass through several other nodes before reaching a customer.

For an idealized geostationary example, take two vertical legs of 35,786 kilometers: ground to satellite, then satellite to ground. Their combined 71,572 kilometers imply roughly 239 milliseconds of light-travel time in vacuum. A request-and-reply exchange following comparable paths starts around 477 milliseconds before processing or network delays.

These are geometric teaching estimates, not a measured service latency. Real ground locations and routes generally make the path longer. They show why quoting only the satellite's altitude can understate the delay experienced by an application.

A long-running batch calculation might barely notice that delay. A rapid back-and-forth interaction might notice it immediately. Orbital siting cannot be separated from the behavior of the service.

Choose an exit as well as an entrance

An orbit's usefulness includes what happens when operations end. A platform that is healthy can execute a planned disposal maneuver. A platform that has lost power, propulsion, or command access may not be able to follow the same plan.

The designer should consider both cases from the beginning. Reserving propellant is useful only if the remaining systems can use it when needed. Relying on atmospheric drag requires an orbit-specific assessment rather than the broad assurance that everything eventually falls down.

Likewise, moving equipment to another orbit is not equivalent to making it disappear. The destination must be evaluated for long-term interactions with other objects and missions.

These choices can affect the original design. A lower operating location may create one set of lifetime and drag costs; a more persistent orbit may create greater end-of-life obligations. There is no reason to postpone this comparison until the revenue-producing portion of the mission has finished.

The best address is one the platform can use responsibly throughout its life. Choosing the route home—or another credible end state—is part of choosing where to work.

Compare complete operating years

Imagine two proposed orbits. One provides more sunlight but fewer useful communications opportunities. The other offers better customer access but requires batteries. The decision should compare a full modeled year, including seasonal extremes and failure cases.

Measure useful jobs completed, not just hours in sunshine. Count maneuvers, network availability, hardware lifetime, and the end-of-mission plan. An excellent first week can conceal a poor annual operating pattern.

What would prove this?

Publish an orbit-specific simulation with stated assumptions, then test its predictions against flight observations. Report energy, thermal behavior, communications access, and disposal requirements together.

A successful orbital data center will not occupy the most glamorous address. It will occupy an address whose advantages match the service it actually sells.

Previous article · Next article