As researchers interested in orbital capacity, it’s surreal to wake up and find Elon Musk commenting on the question that has been central to your work: how many satellites can we fit in low Earth orbit (LEO)? According to a recent interview with the Financial Times, Mr. Musk’s stance is that tens of billions of satellites can coexist in LEO. While we agree this is an important question to ask, especially for someone planning on launching thousands of his own satellites in the near future, his estimation is overly optimistic. In the immortal words of Douglas Adams (and Mr. Musk himself), space is big. But LEO is not big enough to safely accommodate this kind of orbital demand.
Currently, there are more than 4,500 active satellites in orbit. While it took more than five decades to reach a thousand simultaneous active satellites, the growth in the active orbital population has exploded over the last decade, driven largely by companies like SpaceX designing satellite constellations to provide internet access. More than 100,000 more satellites have been proposed, with nearly 40,000 proposed to the U.S. Federal Communications Commission in November 2021 alone. While not all these proposed constellations will move forward, the ones that do and all future constellations will still need to find space in orbit and available spectrum to communicate (which also constrains orbital capacity).
There are two main ways you can think about orbital capacity. One is a probabilistic approach, where orbital capacity is subject to a certain acceptable amount of collision risk under various assumptions. The other is deterministic, in terms of how many satellites can be accommodated under a system of coordinated satellite locations (or slots) that avoids the risk of active on-station satellites crashing into one another. Both frameworks are necessary for different purposes.
Our work focuses on deterministic orbital capacity, which is particularly well-suited to answer capacity questions for large constellations, like how many satellites can fit in LEO and how different constellation designs impact available space for others. We break this problem into two pieces: how to place slots at each altitude in near-spherical volumes or shells and how to stack these shells at nearby altitudes.
Most large constellations (including those built by SpaceX) are designed to ensure that satellites within one of their parked orbital planes will not pose a collision hazard to their other satellites at the same altitude. It’s as if you had multiple intersecting highways, but the cars on each road were precisely spaced to ensure that no accidents would occur at intersections, even while everyone passes through each intersection at full speed. While designing shells to avoid these intra-shell collisions won’t save you from needing to move to avoid debris or shell-transiting satellites, it significantly simplifies operational complexity and improves orbital safety for all, so it’s a smart practice.
With this in mind, looking at the physical volume occupied by a satellite as a capacity metric is misleading – it’s like trying to estimate the capacity of a highway by figuring out how many stopped cars could fit on the pavement. In practice, both cars and satellites moving at high velocities need to leave some following distance between vehicles for safety. On Earth, the size of your following distance depends on factors like the quality of your brakes and tires, visibility, speed, and the risk tolerance of the driver. In space, following distance depends on how well you know the actual locations of the satellites involved, their maneuver capabilities and strategy, and the way the satellites are slotted within their shell. Because orbital velocity is an inherent property of a chosen orbit, extra traffic not designed for compatibility does not result in low speeds and gridlock; it results in dangerous close approaches and sharply increased collision risk that necessitates sophisticated monitoring and periodic collision avoidance maneuvers.
Working closely with research collaborators, our work examines the empirical and analytic limits of constellation size for given minimum separation distances. Using reasonable minimum separation distances, individual shells can fit hundreds to thousands of satellites. Current constellation design practices typically involve individual shells for each operator separated by altitude. Still, with appropriate technical interchange and coordination, it is fundamentally feasible for multiple operators to occupy slots in a shared shell.
Another important consideration for capacity is how closely shells can be stacked vertically. Satellites don’t move in perfectly circular orbits — their orbits are affected by Earth’s non-spherical gravity field and by forces like atmospheric drag and solar radiation pressure. While deviations within the shell can be addressed through minimum separation distances and satellite control strategies, large constellations should be vertically separated from one another to ensure ongoing orbital safety. Naturally, the more altitude variation required by the orbit and control strategy selected by an operator, the fewer shells (and therefore satellites) can fit in a particular region of space. Operators should minimize the vertical space occupied by each of their shells, and regulators should be willing to push back if operators claim overly generous shell widths.
Rigorous, technical work on orbital capacity and a deep understanding of the tradeoffs of various constellation designs are critical to support policy and regulatory discussions about the rational, equitable, efficient, and economical use of LEO. Space is a communal and finite resource — for better or for worse, when one nation uses it, it limits the orbits that can be used by others. We encourage all operators, like Mr. Musk, to factor orbital efficiency into their design process to achieve self-safe and neighbor-safe orbits and constellations in order to maximize the availability of precious LEO volume for all users. Without careful planning, and if we ignore the opportunity cost of planned orbital allocations, there may not be a good place in LEO for tomorrow’s Starlink.
Miles Lifson is a doctoral student at the Massachusetts Institute of Technology and a research assistant for the Astrodynamics, space Robotics, and Controls Laboratory (ARCLab). Richard Linares is the Charles Stark Draper assistant professor of aeronautics and astronautics and the director of ARCLab. Research in ARCLab is funded by various sponsors, including the Aerospace Corporation, Air Force Office of Scientific Research, Air Force Research Laboratory, DARPA, MathWorks, NASA, NASA Jet Propulsion Laboratory, National Science Foundation, and U.S. Space Force. Lifson and Linares have provided paid technical consulting for Amazon on space sustainability and orbit design. The opinions expressed in this piece are their own and do not reflect the views of MIT, our research sponsors, or any other organizations with which they are affiliated.