On October 1, 2026, WilmerHale published a legal analysis of orbital computing that examined how space-based data centers would interact with licensing, contracts, insurance, and data regulation. Orbital data center risk arises when computing hardware, spacecraft operations, and customer obligations depend on one another. A computer failure can become a service interruption, a contractual dispute, or a financing problem, even when the spacecraft remains intact.
The analysis is a law firm’s client alert, not a regulatory ruling or an independent demonstration of commercial performance. Its most useful contribution is the attention it directs to responsibility across the project. A launch provider, spacecraft operator, computing supplier, network provider, and customer may each control part of the service. Their agreements must establish how those parts fit together before a failure exposes a gap between promised performance and available remedies.
The engineering evidence requires equal care. An April 28, 2026 GAO assessment distinguished smaller systems that process information generated in space from large orbital facilities intended to compete with terrestrial data centers. It described substantial technical and economic obstacles to the latter. A successful computing experiment should be evaluated for the workload it actually performed, rather than treated as proof that an Earth-serving facility can operate reliably and profitably at much greater scale.
Heat rejection illustrates the problem. Space offers a vacuum, but that does not automatically remove heat from busy processors. Hardware must transfer heat to surfaces that can radiate it away. Power generation, cooling equipment, communications, and computing capacity consequently have to be designed as an integrated system. The GAO assessment also identified radiation exposure and the difficulty of repair as concerns. These constraints affect usable capacity as well as the cost of placing equipment in orbit.
New Space Economy’s discussion of orbital data-center failure modes provides context for that interdependence. Loss of a connection can make healthy processors unavailable to customers. A cooling limitation can force reduced computing output. Radiation can affect electronics or stored information. These are distinct failure mechanisms, and an agreement covering only complete spacecraft loss would leave important service problems outside its central definition of failure.
The first contractual task is to define the purchased outcome. Installed processing capacity and usable computing service are different measures. A customer may care about completed jobs, response time, data integrity, or the ability to retrieve results. Performance commitments should identify the relevant measure, how it will be observed, and which interruptions count against it. Otherwise, the operator and customer could reasonably evaluate the same mission using incompatible definitions of success.
Responsibility should then follow the parts of the system each participant controls. This is an analytical principle, not a universal legal rule. A computing supplier might control hardware qualification, but have little influence over launch timing or a ground-station outage. The service operator may manage several suppliers and still remain responsible to its customer. Contracts can allocate these exposures differently, so the practical question is whether the customer-facing promise has support throughout the supply chain.
Workload selection changes the exposure. Processing data produced by another spacecraft can avoid sending every raw observation to Earth, but a customer-facing application introduces a different dependency on timely communication. The appropriate reliability standard should follow the intended use. A provider’s successful test for one workload does not establish that another workload can tolerate the same delays, interruptions, or recovery process.
Remedies need similar scrutiny. Compensation for an interruption may be limited to service credits, meaning a discount against future use, rather than payment for every consequence of lost computing access. Whether that arrangement is acceptable depends on the workload and the actual contract. A recoverable batch-processing delay and loss of irreplaceable information have different implications. Customers should assess the cost of maintaining an alternative service alongside the orbital one rather than assuming a remedy restores the interrupted operation.
Insurance cannot be presumed to fill every contractual gap. WilmerHale advises examining availability, exclusions, and policy limits before finalizing indemnities, which are agreements to cover specified losses. A useful analysis distinguishes damage to the spacecraft from loss of data, impaired computing performance, and downstream customer losses. Those categories need not receive identical treatment. An insurance label alone does not establish that a particular scenario is covered or that the available payment would meet the corresponding obligation.
Orbit also does not erase public responsibility. Under the Outer Space Treaty, nongovernmental space activities require authorization and continuing supervision by the appropriate state party. The treaty also preserves jurisdiction and control for the state of registry. These principles provide a reason to reject the assumption that computing beyond national territory operates outside legal oversight. Specific licensing requirements still depend on the mission, its activities, and the jurisdictions involved.
Data-access obligations can follow the provider as well as the equipment. U.S. law on provider-controlled records requires covered providers to comply with specified preservation and disclosure obligations for information in their possession, custody, or control, regardless of whether it is located inside or outside the United States. That does not create unrestricted government access to every orbital dataset. It does show why physical location alone cannot establish freedom from applicable disclosure duties.
Mission retirement belongs in the same analysis as service delivery. Replacement, transfer, disposal, and the customer’s ability to recover information should be addressed before hardware reaches the end of useful operation. A financing model that counts service revenue without accounting for retirement responsibilities would leave part of the mission unexplained. Agreements also need to distinguish customer access to data from ownership of the physical equipment carrying it.
The resulting business test is broader than whether a processor can run in orbit. A credible provider needs a supported operating model and a chain of obligations that remains workable through degraded service, supplier failure, and mission retirement. Clear contracts cannot eliminate engineering uncertainty, and a demonstration cannot settle every liability question. They can make the remaining uncertainty visible enough for customers, insurers, and investors to decide which risks they are prepared to accept and what evidence would justify a larger commitment.
Useful Books Available on Amazon
