Gravel2Gavel Construction & Real Estate Law Blog
Data centers … in space.
How did the “final frontier” become a genuine consideration for siting and constructing next-generation data centers? Perhaps it is the inevitable result of demand greatly outstripping supply in two of the great pillars of the data center ecosystem: real estate and power supply. As Pillsbury recently reported in the Pillsbury Guide to Data Centers, some 11,800 data centers were reported to already be operating worldwide in 2024. Yet the demand for centers is expected to rise by as much as 22% annually until 2030, placing significant constraints on the ability of operators to locate sufficient real estate to build and operate them all, obtain the necessary permits for construction and operation on a timely basis, and, significantly, ensure the availability and reliability of electricity to power and cool the components. Or perhaps growing concerns over security, resiliency and environmental impacts are driving operators and users to look for alternative solutions.
What is indisputable is that players across the data center ecosystem are searching for new solutions that can simultaneously help meet the exponentially growing demand for data volume and circumvent the increasing challenges facing expanding terrestrial infrastructure. While some players are looking seaward—to the harbors, shores and even the ocean deep as future sites of permanent or mobile solutions—others are looking up. The 21st Century’s commercial space renaissance has led to orbital and extraterrestrial solutions emerging as legitimate—and arguably frontrunning—contenders to resolve the data center dilemma.
Powering Up
Access to consistent power is the lifeblood of any data center. It is the top entry in any site selection checklist—the gating factor in determining whether, when and where a data center will be built. The power source’s vitality is twofold—it enables the main components to operate, and it maintains temperature control to prevent overheating.
Data centers are approaching and exceeding gigawatt-magnitude loads and demand exceptional reliability—up to 99.999% online service—due to the computational and financial costs of power interruptions. The massive power demands and strain on the energy grid have played a key role in more restrictions being placed on siting. Even Loudon County, Va., home to Data Center Alley, has recently eliminated the “by-right” approval process for new data center development due to the strains the operations are placing on the county’s energy infrastructure.
Developers have responded by exploring alternative power sources beyond those sources commonly deployed in regional or national energy grids. Nuclear, solar and other renewable energy sources have shown the most promise, particularly as operators look to balance their energy needs against growing environmental impact concerns that the massive upscale in centers and gigawatt hours attract. As the demand for zero-emission output from international and hyperscaler players increases, the lure of accessing near-constant, unfiltered solar energy makes space an attractive destination for future infrastructure expansion.
The Real World of Off-World Infrastructure
Moving data centers off-planet sounds far-fetched, until one recognizes that the baseline infrastructures needed to establish incremental and scalable networks already exist. Satellites, in-orbit laboratories and lunar modules (such as lunar landers) provide the foundational architecture for data storage and processing in space. These platforms employ a combination of radiofrequency and optical laser links to transmit and receive data both to and from the Earth, as well as intra- and inter-networking in space. By transmitting data among spacecraft, networks can relay data around the globe at rates that cannot be matched by terrestrial telecommunications infrastructure. Pioneering companies currently working on space-based data solutions have started from these first principles and are currently developing and deploying initial data centers in the form of low Earth orbit (LEO) satellite networks (Starcloud and Axiom Space), and modular lunar storage (Lonestar Data Holdings).
Aided by the increasingly lower costs and emerging competition for commercial launch services, companies will soon be able to deploy, scale and iterate on technology platforms at essentially the same rate as terrestrial systems. Lower costs to orbit have also facilitated the development and advancement of space-based services—known as in-space servicing, assembly and manufacturing (ISAMs)—which seek to enable more complex architectures for the space environment through the launch of raw materials and modular units for in-situ fabrication and construction (for example by companies such as Redwire or Spacedock).
Unlike terrestrial infrastructure, nearly all space-based platforms are designed to rely on solar power as a principal source of energy over the course of the asset’s operational lifetime. In this environment, sunlight is nearly constant for most platforms and without the Earth’s protective atmosphere operating as a filter, this energy is delivered with high intensity (measured in watts per square meter). (See also Pillsbury’s article on Microwave Transmission of Solar Energy from Space.) Spacecraft will orient themselves or their solar arrays towards the sun as the platform or celestial body orbits around the Earth. These arrays are used to charge and recharge onboard battery cells (commonly lithium-ion) that are relied upon to power vehicle operations, including transmissions and maneuvering, and to power the spacecraft while operating in the Earth’s shadow.
This dependence on solar energy is also true for space-based platforms that will operate from the surface of a celestial body, such as the Moon. Both countries and companies have keyed in on ideal lunar sites for permanent and semi-permanent infrastructure installations based on the availability of near-constant exposure to sunlight. Illuminated up to 88% and receiving sunlight up to 92% of the time on average, the Peaks of Eternal Light along the Hinshelwood, Peary and Whipple craters at the lunar north pole, and the Shackleton and De Gerlache craters at the lunar south pole, are notable locations that space agencies and companies identified as landing targets for past and upcoming lunar missions.
The expansiveness and remoteness of space has also enabled nuclear-powered technologies to garner political and commercial support for resurgence at levels not mirrored for Earth-bound facilities. This is evidenced by the Trump administration’s recently released Directive on Fission Surface Power (FSP) Development, identifying nuclear as “both an essential and sustainable segment of the lunar and Mars power architectures.” NASA has already released a draft Announcement for Partnership Proposals seeking to “make power available on the lunar surface and to energize the space industrial base to support a future lunar economy through the deployment of a (100 kW) FSP system on the lunar surface.” NASA and DOE’s Fission Surface Power System (FSPS) has also previously awarded three $5 million contracts for the design, cost estimation and development schedule of an up to four-by-six meter, 6,000-metric-ton fission reactor capable of delivering at least 40 kW of continuous electric power over 10 years by 2030. Though the electrical output of a single FSPS reactor might only satisfy modest data center demands, such reactors could be scaled to meet the growing load demands of data centers and other lunar infrastructure. More importantly, these initiatives signal the government’s commitment to partner with industry in the development and advancement of nuclear technologies for sustainable space infrastructure.
Security and Resilience
In-space data centers also provide solutions for players looking for added resiliency, security and data sovereignty. As described in the preceding section, most space-based systems provide a physically diverse pipeline through which customer data can travel, apart from terrestrial telecommunications infrastructure. While some satellite networks intentionally select ground stations that are collocated with data centers (AWS Ground Station), the space-based data center paradigm would provide the reverse arrangement for purposes of resilient or redundant capacity.
Physical separation also enhances the security of the assets, making them harder to access and reduces the likelihood of intrusion. For much of the data stored on-orbit, particularly in celestial modules, the data would be in a form of long-term or backup storage, where users are not accessing the information on a constant basis and the modules function as a sort of “vault.”
As the nationality of the spacecraft is established based on the licensing authority or the launching state of the object, these platforms also offer a unique solution for data sovereignty. As countries adopt additional requirements for companies and individuals to hold certain information within the country of origin, space-based data centers offer additional locations in which data can be stored within the dictates of the regulations.
High Risk, High Reward
If space-based data centers are essentially an adaptation or extension of existing technologies, why have they not yet been deployed en masse? Well, as the cliché goes, space is hard.
For starters, while the foundational space platforms exist, the data center technology must still be adapted for the harsh environments of space. Adapted or new technologies for performing the computing functions of data centers must be developed and tested before being deployed into the space ecosystem. Once launched, the platform itself can typically only be modified or repaired in minimal ways, such as with software updates. Physical changes to the platform would require the operator to build and launch a replacement spacecraft. While ISAMs may introduce these capabilities in the future, today, many defects will result in the premature end or severe diminishment of the spacecraft.
Costs of accessing space have plummeted over the last decade, but building and sustaining infrastructure projects in space are still massively expensive, capital-intensive endeavors. Beginning with the upfront costs—the minimum-viable network must be designed, built, launched and commissioned before any revenue can be earned on services—the operators must also plan for ongoing operational costs and replacement vehicles. In low Earth orbit, spacecraft are typically designed for operational lifetimes of five to eight years before they are deorbited and a new spacecraft launched as a replacement. On the lunar surface, the assets will likely be hardened to operate for longer periods of time (using geostationary orbit as a benchmark, 15 – 20 years), but its trade-off is that there is less opportunity to iterate on the technology so each future module must also be backwards compatible to interface with existing infrastructure. There are also associated costs for deploying (or leasing) and operating the ground network that communicates with the in-space assets, and which are situated in a multitude of jurisdictions all across the world.
Operating in space carries its own risks, from initial launch through deorbit. While SpaceX’s Falcon 9 is beginning to make launch look routine, generally launch still has a distance to go before it becomes as routine and reliable as land and air cargo transport. Competition is also relatively nascent, so ensuring every operator that wants regular launch services can secure the volume they need may be difficult over time unless the launch sector right-sizes at the same pace.
Once on-orbit, operators must navigate additional challenges. As mentioned above, objects in space cannot be repaired in the same way they can on Earth. Space assets must also monitor for and avoid collisions with other objects, including other satellites and stations, as well as debris and meteorites. Spacecraft must also be hardened to protect against radiation and other space weather. The very same solar radiation that powers the solar arrays can cause serious damage. Fluctuations of solar wind and geomagnetic storms can corrupt software, cause short-circuiting and reduce solar cell output.
Despite the extreme cold of space, maintaining a regulated cooling system in the sealed environment of a data center may also introduce its own challenges. The air-based cooling employed in terrestrial data centers is not an option in the vacuum of space. The heat generated by operating centers would have to be ejected via suitably designed large radiators that could still maintain the integrity of the space station’s structure. The cooling system would also have to adapt to rapidly changing temperatures based on the center’s immediate level of sun exposure. And while liquid-based cooling systems have increased in popularity here on Earth, it is unclear whether similar systems would have equal effect in the lunar context, considering the need to account for the potential impacts of lower gravity environments on fluid behaviors and the associated costs of transporting coolants from Earth to the lunar site.
Finally, there is still some regulatory uncertainty around operating data centers in space. While the regulations around the licensing of satellite networks are fairly well established (though different in every country), there is less certainty and consistency around ISAMs and use of lunar and celestial resources. In the United States, the Trump administration recently released an Executive Order instructing the Commerce Secretary to propose new regulations surrounding all novel space activities not currently covered by existing regulations. While orbital data centers arguably rightly fall under the FCC’s licensing regime for communications satellites, the case is less clear for modules operating on the lunar surface. Additionally, as we previously covered in Pillsbury’s article on Lunar Natural Resources, the ability to deploy permanent and semi-permanent structures on the lunar surface is also a matter of international legal debate and the provenance of the first Trump administration’s successful Artemis Accords.
A New Space Race?
With all these considerations in mind, at least three separate organizations have started or are nearing the stage where they can demonstrate proof of concept.
- Lonestar Data: With its sights set on the Moon, Lonestar is the first to launch and demonstrate an initial proof of concept. Earlier in 2025, the company sent a data-carrying device the size of a shoebox to the Moon as part of Intuitive Machines’ Athena Lunar Lander to test its ability to upload, download and transfer data. Despite the failure of the Intuitive Machines lander, the Lonestar device successfully completed several of its test objectives. The company plans to launch a data center to orbit the Moon in 2027, and plan to establish data centers on the lunar surface.
- Starcloud: Starcloud is working to build a network of low Earth orbit data center satellites, powered by a grid of solar panels that would extend approximately two and a half miles wide. Tests for its data center network will begin with satellite launches in 2025 and 2026. Its goal is to have a network up and running by the early 2030s, deploying 40-megawatt orbital data centers powered by solar energy.
- Axiom Space: On top of attempting to create the first commercial space station (which will have a data center onboard), Axiom aims to create its own network of orbital data center nodes in low Earth orbit, beginning with launches scheduled for late 2025. These nodes are intended to provide secure, scalable and cloud-enabled data storage and processing, and artificial intelligence/machine learning solutions directly to satellites and other spacecraft with the capability to operate independently of terrestrial infrastructure.
Thus, what may have sounded like science fiction when you started reading this article may actually be coming to an orbital arc near you in just a few short years.
RELATED ARTICLES
Microwave Transmission of Space-Based Solar Power: The Focus of New Attention