If you read my previous article on the Fleet Maintenance Paradigm for Satellite Servicing, you might be wondering what sort of things Fleet Maintenance can be used for? While in theory you can do a very wide range of modifications to a satellite using a servicer1, some applications lend themselves better to Fleet Maintenance than others. Here are a few of my thoughts on high-payoff applications for Fleet Maintenance:
Compute Refresh
Using modular interfaces, Fleet Maintenance servicers can periodically update the on-board computing capabilities of a constellation.
Even in a world that’s no longer strictly following Moore’s law, computing systems still tend to get consistently better and more capable over time, both in terms of raw computational capability as well as computational efficiency. As on-board computers on spacecraft become more powerful, they enable many new capabilities such as sensor fusion and image processing, communications optimization, and increased autonomy.
While edge computing processors are unlikely to wear out in the 5-6 year lifespan of a typical LEO constellation satellite, they do become rapidly outdated. It’s worth noting that over the multiple years it takes to build and launch a given constellation generation, it’s not uncommon to do one or more design iterations or technology refreshes. For instance, while I could be wrong, I’d be kind of surprised if a Starlink satellite rolling off the assembly line in Feb 2026 was using the same processor model on-board as one rolling of the line in Feb 2024 or 2022.
If it’s worth updating processors periodically at the factory, it can potentially be worth updating processor modules on-orbit if you’re already planning regular visits to each satellite in your constellation.
SDA Sensors and Defensive Countermeasures
If you’re a military or even remotely dual-use spacecraft, congratulations, you’re probably now a target2! With hostile powers rapidly developing ground-based anti-satellite and orbital counterspace capabilities3, it is becoming increasingly imperative for new military and dual-use spacecraft to start considering defensive options.
Some companies4 are developing the spacecraft equivalence of security cameras you can mount on your satellite to detect hostile spacecraft approaching. There are also a range of potential defensive countermeasures you could potentially mount on a spacecraft, things ranging from RF jammers, optical/IR dazzlers, decoys, deployable small sentinel/harassing spacecraft, etc. Basically, you want the ability to detect hostile spacecraft as they approach, and prevent them from damaging your spacecraft. What defensive capabilities you’d want to add to a space asset varies, based on how high of priority of a target you think you might be. But in the coming era of peer competition, being defenseless is probably a bad idea.
So, a second area for Fleet Maintenance will be delivering and installing or exchanging SDA5 sensors and defensive countermeasures. If you’ve been watching the development of drone warfare during the current war in Ukraine, the iteration speed of measure and countermeasure can sometimes be measured in weeks. Having 5 year old countermeasures on your spacecraft may be functionally equivalent to being defenseless. But Fleet Maintenance gives you a way of frequently and affordably updating your constellation’s ability to defend itself from hostile actions.
High-Rate Comms
Most commercial, civil, and military satellites use high-rate communication systems to send data either down to Earth or between satellites in a network. This is a rapidly evolving field, especially in the realm of optical intersatellite links6, and Fleet Maintenance gives you a way to keep your constellation up-to-date with the latest and greatest communications systems.
While I had to speculate about constellation operators updating their compute processors throughout the buildout of a constellation, periodic upgrades of intersatellite links are something clearly discussed in the public literature. For instance, SpaceX with Starlink launched their initial V1.0 satellites without optical intersatellite links, their early V1.5 Starlink satellites had a basic OISL system7, and it was upgraded at least once more with the v2 Mini configuration that is currently being built and launched. While most of the early Starlink satellites that lack OISLs have now been retired8, something like 1/3 of their fleet is currently using the first-generation optical links.
Especially with the rapid evolution of in-space laser communications networks, being able to upgrade your comms solution, whether you’re a customer satellite or operating nodes in the network, can be benificial for increasing data throughput. With the right design of interface ports and data network design within the satellite9, these types of high-rate comms systems can be modularly upgraded one or more times over the lifetime of a given space asset within a constellation using Fleet Maintenance servicing. Even upgrading hundreds or thousands of satellites in a constellation is not necessarily impractical if you have a small constellation of servicers visiting each satellite on a regular cadence.
Batteries
Low Earth Orbit is one of the more brutal regimes for high-power lithium-ion batteries, with satellites seeing 5000-6000 day/night cycles per year, or 30,000 to 35,000 over the course of a typical LEO satellite’s 5-6 year lifetime. Current lithium ion batteries degrade significantly over a large number of cycles meaning that you typically have to only use a fraction (typically ~25%) of the battery’s theoretical or nameplate energy capacity. For most satellites this is just a nuisance, but for some high-power applications like Synthetic Aperture Radar satellites, if you knew you could swap batteries every year or two, you wouldn’t have to baby the batteries as much and could use a much more aggressive and deeper depth of discharge and/or discharge rate. Because battery degradation is easy to measure and predict, this is precisely the kind of pre-planned maintenance that the Fleet Maintenance paradigm works well for.
Hosted Payloads
Particularly for science or communications satellites, there is often value to hosting secondary payloads from other organizations or companies10. The challenge is that hosted payload developers/operators, by their nature, tend to have more marginal business plans than those operating full constellations. This increases their odds of not being able to get the financing they need on time, which might cause them to have to back out on a planned host. With traditional spacecraft approaches, there’s really only one time to integrate a hosted payload, and once that’s past, you’ve missed your opportunity. So, while they can be quite lucrative, they’re also potentially a high-risk/high-payoff proposition for a satellite operator.
With satellite servicing in a Fleet Maintenance paradigm, you now have regular and affordable opportunities to revisit a host satellite even after it has been launched. This means that in a world where one or more constellations was outfitted with modular hosted-payload interface ports11, it would be possible to keep a catalog of potential host spacecraft, including data on their available resources (number of open ports, spare power, available comms bandwidth, etc.), allowing hosted payload developers to select the optimal constellation to install their payloads onto, and deliver and install those payloads via Fleet Maintenance when the servicers next visit the desired host vehicles.
While currently there are some so-called “condosats” that are designed to host a large number of payloads12, there’s always a transition point where condosats stop being an option. You can test out a sensor or other payload that wants to be in a constellation first on a single satellite or pair of satellites, but at some point, you have to bite the bullet and build out a constellation, or work a hosted payload deal with someone else’s constellation, and hope the relative timing of your two businesses stay aligned. But with satellites designed with in-space swappable hosted payload interfaces, and Fleet Maintenance servicers, condosats and other payload hosting spacecraft become much more like terrestrial condos and real estate — you can not only move new customers in at many points throughout their lifetime, but you can also remove customers who no longer need hosting. Could you imagine trying to run a Condominium on Earth if you could only move people in once, and you could never move them out until the building was ready to be torn down? Fleet Maintenance fundamentally unlocks an entirely new world for hosted payload constellations and host vehicles.
Bandaids for Common-Mode Failures
If you ask any new constellation developer what one of their biggest fears is, they’ll likely say common-mode failures that take a little bit of time to manifest13. Imagine being a year or two into building out a new constellation only to find that all of your satellites to-date share a common flaw that dramatically shortens their lifespan or cripples their functionality? Right now, your only solution is prayer and insurance.
While not all common mode failures can be fixed with some sort of “band-aid” module plugged into an interface port on the satellites, a lot of them potentially can. I’ll get into modularity strategy in future blog posts, but this part of the reason why I’d always make sure you have at least one general-purpose port that can be used for such band-aids. While sometimes you’ll be repairing just one satellite with a specific problem, if you do get unlucky enough to have a common-mode failure, having a way to patch-up several satellites rather than having to write them off would be a welcome alternative for most operators.
Also, if you know you have a way of patching common-mode failures, it makes it easier to take more risks in satellite design. Basically, if you know you can get a bandaid there fast enough, you can treat your Fleet Maintenance servicing as your B-string (or at least as your second layer of fault tolerance).
Refueling
You’re probably wondering why I didn’t list this first. I absolutely think refueling is a critical servicing capability, and with Space Force interest in the capability14, it will likely be one of the next ones fielded operationally.
My initial main reason for listing this later is that the space vehicles most likely to use refueling15 are also the ones that definitionally don’t stay put in anything resembling an orderly constellation pattern, so they might be a better fit for the second servicing paradigm (Multi-Client Opportunistic Servicing). However, after thinking about this more, these types of spacecraft, while they likely won’t be flying in a nice and stable formation, like a communications constellation or earth observation constellation, will likely be operating in and near such constellations. So, they might still be able to take advantage of Fleet Maintenance servicing16.
Other types of spacecraft that are less fundamentally dynamic in nature17 are less likely to do a ton of maneuvering in nominal operations. They typically mostly focus propulsion on orbit-raising to reach operational orbit, stationkeeping/collision avoidance maneuvers when in their formation, and then post-mission disposal. But especially for defense and dual-use spacecraft, operating in an increasingly contested environment, having the luxury of being able to exchange fuel reserved for disposal for evasive maneuvers during a conflict might be a life saver. Especially if you know that you’ll have a regular Fleet Maintenance top-up opportunity before long.
Additionally, for satellites operating in high orbit, most of their propellant is used for orbit raising and end of life disposal. If you know you can count on periodic Fleet Maintenance top-ups, you might be able to launch with a smaller tank rather than trying to carry all of your propellant from day one. This can lower your up-front cost of fielding a constellation by non-trivial amounts by allowing you to cram a few extra satellites on each launch, while allowing you to buy propellant later, potentially at a lower price, if launch competition with new entrants like Stoke, RocketLab, and Firefly is successful.
Summary/Observations
Hopefully these examples helped start you thinking about potential ways to use this new paradigm of servicing. The whole idea of being able to count on regular, pre-scheduled, and super-affordable servicing starts moving constellations out of the mentality we’ve been stuck in for most of the Space Age, where most satellites were alone forever after launch. While not all of these examples will apply to each class of satellites, many will.
Some key themes looking back are that Fleet Maintenance works great for:
- Things that go obsolete or wear out faster than the typical lifetime of your constellation, especially things that go obsolete driven by commercial or military competitors.
- Things that are value-dense and fit well into small volume modular packages that can be readily installed using standardized interfaces18.
Also, hopefully this has helped show how even short-lived, mass-produced LEO satellites with 5-6yr lifespans can benefit from satellite servicing. You don’t need to go all the way to a Spacecraft of Theseus level of modularlity and immortality19 for servicing to be economically relevant.
- A DARPA program I was involved with in the early 2010s was looking at the idea of hacking off the deployed reflector from a dead GEO satellite and building a new satellite around that reflector using a bunch of satellite lego bricks. The limits of what you can achieve with satellite servicing are only bounded by “the height of your creativity, or the depth of your neurological disorders” as a friend once put it. But this article is focused on trying to highlight some of the more practical things you can do. ↩︎
- I mean, I guess if your satellite has no real economic or military utility, it might not be a target. And the less useful you are, the lower priority you probably are for hostile powers. But if your satellite is even remotely useful, you’re probably somewhere on someone’s target list. ↩︎
- Orbital counterspace meaning spacecraft designed to harass, impede, neutralize, or even destroy other spacecraft. ↩︎
- Scout Space, Katalyst Space Technologies, and others. ↩︎
- Space Domain Awareness – which is defined as “the capability to detect, track, identify and characterize Space objects and the Space environment” according one reference. ↩︎
- Basically laser-based communications systems (“OISLs” or Optical Inter-Satellite Links) that work in a point-to-point fashion. The current record is 200Gb/s, achieved by a NASA/MIT payload called TBIRD in 2023. ↩︎
- Google AI was claiming a 10-20Gbps data rate for the V1.5 OISLs vs 100+ Gbps for the v2 minis. ↩︎
- Only about ~374 are still on orbit as of end of Feb 2026, according to a Google AI search. This represents about 4% of the current Starlink fleet. ↩︎
- To take advantage of upgraded OISLs, you’d likely want to design the network within the satellite to handle a much higher data rate than the version of OISLs the satellite launches with. For a typical short-lived LEO satellite, most networks will likely have some predicted performance levels over that 5-6 year span, allowing satellite designers to properly size the throughput inside the satellite and through the modular connector to the OISLs. The goal would be to have the OISL be the comms bottleneck, not the rest of the system. ↩︎
- I don’t know if this is still the case, but in a customer discovery interview I did with Matt Desch of Iridium back in 3 or 4 years, he pointed out that almost 10% of Iridium’s profit came from just a single hosted payload they were flying on their second generation satellites. IIRC it was some sort of aviation-related payload. ↩︎
- It’s a discussion for another day, but the plan I was taking to try and crack the chicken-and-egg problem of “who goes first, the servicers or the interfaces?” was to work with next generation comms constellation developers to incorporate servicing interfaces as standardized hosted payload interface ports. They didn’t have to be convinced that servicers will ever exist for it to make sense to put hosted payload interfaces on, and once they’re there, you now have a proven real estate market just looking for delivery vehicles. They don’t have to believe in servicers at the start, but once they exist, they will happily sell additional capacity on their fleet, since that’s now basically pure profit for them.
Unfortunately, the timing didn’t work out — I started trying to push this right as Iridium and OneWeb were wrapping up their most recent constellations, and their replacements were at least 5-10yrs out, and Starlink and Kuiper didn’t like hosted payloads. Selling to Voyager and having them merge Altius out of existence didn’t help either. But I think there’s still a real oppty here for someone willing to make the right kind of modular interface. As I said though, more for a future post. ↩︎ - The most prominent example I know of is my friends at Loft Orbital/Loft Federal, which has flown at least a half dozen “YAM” (Yet Another Mission) hosted spacecraft platforms, with something like another two dozen or more in their backlog. ↩︎
- Not that this has ever happened to satellite operators. Like take Worldview-4, a Maxar commercial imaging satellite that failed 2yrs into its planned 10-12yr lifetime when they discovered that the CMGs they got from a vendor had bearings that wore out prematurely. If this had been a constellation, they could’ve had hundreds of satellites up there, all with the defective CMGs before enough data started trickling in to notice the problem.
As it is, when that failed, their stock tanked and they had a new CEO within a month. Had Worldview-4 existed in a world of Fleet Maintenance satellite servicing, it might have been a non-event, since Worldview-4 was in a popular orbit where there likely could’ve been servicing spacecraft, even if it wasn’t part of a big constellation itself. Fleet Maintenance makes servicing everything near the primary customer constellation easier and more affordable. ↩︎ - The Space Force is requiring all of their RG-XX next generation GEO recon satellites to be outfitted with refueling ports, and are actively soliciting information on refueling architectures to support their future RG-XX constellation. ↩︎
- Think reconnaissance and orbital warfare satellites and orbital transfer vehicles, and reusable in-space transfer stages. Basically, most satellites gently sip propellants, while rockets guzzle it with reckless abandon. ↩︎
- Either by maneuvering over to the nearest fleet maintenance servicer when they’re running low on fuel, or “throwing an interrupt” and having the Fleet Maintenance servicer visit them before returning to their regular route. ↩︎
- Comms, earth observation, position, navigation, and timing (PN&T), missile warning satellites, etc. want to generally stick close to a specific pattern. They can do some minor evasion and maneuvering if critically important, but for the most part generally want to stay in formation as much as possible. ↩︎
- While you could absolutely design a whole modular solar-array wing, most likely you’ll see a better bang for the buck with high value-density modules like compute, batteries, hosted payloads, and defensive countermeasures. ↩︎
- Though I do think “Persistent Platforms”/”Immortal Satellites” will eventually play a role, especially if we end up talking about building out 100s of GW of space-based data centers in the coming years/decades. But you don’t have to wait for that day for servicing to be relevant to satellite operators. ↩︎