Kilometre-scale orbital data centres may clear debris by avoiding rocket bodies and defunct satellites

The sky above us is becoming crowded. For decades, human activity has left behind a chaotic layer of space junk: spent rocket stages, broken satellites, and countless fragments of metal and paint. This debris travels at hypervelocity speeds, roughly ten times faster than a rifle bullet. At such speeds, even a tiny speck can punch a hole through a spacecraft, while a larger piece can shatter a satellite into thousands of new fragments. This creates a dangerous cycle where collisions generate more debris, which leads to more collisions. Scientists call this the collisional cascade, a runaway process that could eventually make certain orbits unusable for centuries.

Now, a new generation of proposals aims to place massive data centers in this very environment. Companies are planning to build structures in low Earth orbit that span several square kilometers, essentially floating server farms powered by continuous sunlight. These facilities would be enormous, with solar arrays stretching for kilometers, designed to handle the immense computing power needed for artificial intelligence. The critical question for these projects has always been how the space junk would affect the structure. Would a stray piece of debris tear a hole in the delicate solar panels? The new research flips this question on its head. Instead of asking what the debris does to the structure, it asks what the structure does to the debris. The answer reveals a surprising possibility: these giant orbital platforms could either become the worst polluters in history or the most effective cleaners the space has ever seen, depending entirely on how they are managed.

The study, led by researcher Jiawei Jing, simulates the behavior of these kilometer-scale structures over a fifteen-year period. The model tracks two distinct types of objects crossing the path of the data center. The first group consists of tiny fragments and small lethal objects, which are too small to be tracked by ground radar but are heavy enough to destroy a satellite if they hit it. The second group includes the large, intact objects: defunct satellites and rocket bodies that are well-known and tracked. The research finds that the thin solar arrays act very differently on these two groups. When a small, fast-moving fragment hits the lightweight sheet, the impact slows the fragment down significantly. The sheet punches a hole, and the momentum exchange is enough to drop the fragment's orbit, causing it to burn up in the atmosphere within a short time. In this way, the structure naturally sweeps the sky clean of the untrackable, mission-ending debris.

However, the interaction with the large, intact objects tells a different story. When a massive rocket body or dead satellite crosses the path of the thin array, the sheet does not stop it. Instead, the impact acts like a high-speed projectile hitting a wall. The energy deposited is so intense that it can strip away a portion of the large object's surface or, in the worst case, shatter the entire object into a cloud of new debris. The study calculates that if even a tiny fraction of the large object's mass—less than one percent—is stripped away or if the object breaks apart completely, the structure becomes a net source of new space junk. In simulations where the structure does nothing to avoid these large objects, the number of lethal, untrackable fragments in that orbital layer could grow to be three to fifteen times higher than it is today. The structure would effectively be seeding the sky with new hazards, undoing the natural cleaning it performs on the smaller debris.

The solution lies in a specific type of avoidance maneuver. The research shows that if the data center is programmed to dodge only the few hundred large, tracked objects—rocket bodies and defunct satellites—it can remain a net remover of debris. By steering clear of these massive targets, the structure allows the smaller, untrackable fragments to pass through and be slowed down by the solar arrays, while avoiding the catastrophic collisions that would create new clouds of junk. This selective avoidance is surprisingly efficient. The study suggests that if the structure can successfully screen out 93 to 99.6 percent of these large objects, depending on the altitude, it will clean the orbit faster than it pollutes it. This is a much lower threshold than the near-perfect avoidance required if the structure tried to dodge every single piece of tracked debris, including the small fragments it is meant to clean up.

The cost of this selective cleaning is manageable, but it depends on how well we know where the large objects are. To dodge a rocket body, the structure needs to move its massive frame just enough to miss the crossing point. If the position of the rocket body is known with high precision, within a few hundred meters, the fuel required for these maneuvers is relatively small. However, if the position is uncertain by several kilometers, the fuel needed to ensure a safe miss increases dramatically, potentially consuming a significant portion of the structure's mass every year. The study also highlights a critical vulnerability: the dense computer modules inside the structure. Unlike the thin solar panels, these heavy modules are fragile targets. If a large object hits a computer module, it would likely shatter the module, creating a massive new debris cloud. Therefore, the computer modules must be protected by dodging every single tracked object that could destroy them, a much stricter requirement than for the solar arrays.

The findings challenge the current way space agencies and companies assess risk. Standard safety rules focus on the probability of a collision damaging the spacecraft. For these giant structures, that approach is insufficient. The study argues that regulators and operators must look at the outcome of the crossing itself. If a large object crosses the array and breaks up, the structure has failed its environmental duty, even if the structure itself survives. The research suggests that the design of these future data centers must prioritize the protection of the orbital environment over the protection of the structure itself. By letting the small debris pass through to be cleaned up, while actively dodging the large, tracked junk, these kilometer-scale platforms could become the first human-made objects to actively reduce the population of space debris.

This outcome is not guaranteed. It relies on two factors that have not yet been measured in a real orbital environment: exactly how much mass a large object loses when it grazes a thin sheet at orbital speeds, and how many new fragments that lost mass creates. Until these physical interactions are tested, the exact efficiency required for the data centers to remain net cleaners remains a simulation. Nevertheless, the study provides a clear path forward. It demonstrates that the fate of low Earth orbit in the age of giant space infrastructure is a design choice. If these structures are built to dodge the large, tracked objects and let the small debris pass, they could help clear the sky. If they are built to avoid all tracked objects or, worse, to ignore the large ones entirely, they could trigger a catastrophic increase in space debris, making the very orbits they need to operate in unusable. The difference between a space elevator that cleans the sky and one that fills it with junk is simply whether it knows which objects to dodge.