2025 Week 15 | Satellite Autonomy and Propulsion Open Door to True Space Domain Changes

I. Tracking

1. Automated vs Autonomous: Technology to Ensure Satellites Require Fewer Human Inputs is Developed

In recent months, a breakthrough in satellite autonomy has been demonstrated by startups NOVI and Sedaro, marking a significant step toward fully autonomous space operations. To be clear, satellites are already masters of automated operations; they fly around the planet and do their work 24/7/365, without ever having any hardware maintenance (being in space) and still providing services. Some GPS satellites launched in the 90s worked 30 years non-stop before being retired.

But the innovation coming from the startups allows satellites to analyze data and make decisions on their own, reducing the need for constant communication with ground control.

The introduction of autonomous satellite operations also raises important challenges and risks. One of the biggest concerns is safety. While we’re not completely clear about how automated these satellites are, and here, semantics matter. Are they highly, highly automated (like an aircraft autopilot that can autoland) or are they autonomous (like a new AI system powered by a neural network and vast complex and probabilistic algorithms). Without getting into the complicated discussion of deterministic vs. stochastic systems, it’s worth a closer look to better understand the future of automated/autonomous spacecraft.

If it’s the AI route, then for mere Earthlings with the mass-consumer highly-imperfect AI experiences, there are serious questions that exist in terms of what percentage of decisions made by these satellites will be based on hallucinations or actual, measurable logic. Until recently (and still today to a certain extent), many operators and engineers continue to doubt the use of generative AI to make decision, for the simple reason that finding a root cause to an error is nearly impossible. The network mesh that is used by Gen-AI systems to come to an output means that between input and output, there is a black box that cannot trace back to an error. The known example is the Sun-Moon problem: if you tell an AI to look at the bright object in the sky that measures this size, it may confuse the Sun or the Moon (which to us, are about the same size, just like you see during an eclipse). With non-gen-AI software systems, a series of plain if-then statements can be followed to identify where something went wrong and fix it. For gen-AI used in the likes of consumer ChatGPT, that’s not possible – each output is different, and even if it goes well nine times out of ten, there’s still a 50% chance of failure the tenth time. The consequences of a hallucinating spacecraft could lead to unforeseen consequences, such as accidental collisions. Ensuring that these satellites operate safely and ethically will require careful oversight and further development in technology to handle unexpected situations. Unfortunately, it will most likely require a severe accident for this to happen.

Still, this decision-capability that goes beyond autopilot is promised to streamline satellite operations, enabling more responsive, efficient missions. In particular, for military and commercial applications, autonomous satellites can adjust orbits, manage resources, and optimize communication links based on real-time data, all without waiting for instructions from Earth.

The potential benefits of this technology are vast. With the growing number of satellites in low Earth orbit, particularly in mega-constellations, automation is critical. Satellites can more rapidly adapt to the dynamic and congested environment of space to avoid collisions or interference. Automated operations could also reduce costs (by minimizing staff requirements), improve efficiency (without needing to wait for a contact time from a ground station), and speed up innovation in satellite missions (satellites capable of retasking rapidly). For businesses and defense agencies, this means better performance and quicker deployment, which is key to maintaining a competitive edge.

Despite these challenges, the promise of autonomous space operations is coming, and could pave the way for more advanced, efficient, and scalable space operations and even exploration. Astronauts beware.

2. Propulsion System Startup Aiming to Change the Way Satellite Travel On Orbit

Portal Space Systems recently raised $17.5 million to develop a highly maneuverable spacecraft that doesn’t solely depend on large quantities of propellant. Regarding space innovations, this is a major problem to tackle if we want to evolve our use of space. Currently, humanity remains highly dependent on launching their satellites right into the orbit they want (through a direct launch to the required orbit), or close enough (as a ride share). There are some developments of ‘space tugs’ or ‘space platforms’, such as Blue Origin’s Blue Ring program or the oddly named D-Orbit, where you simply latch onto a provided ‘mothership, but you are still sharing the technologies and are potentially limited by the other payloads sharing the system with you. Unfortunately, once you’re in space, the most common method today to go ‘anywhere’ from your orbit is to carry your own engines and propellant. Lots of it. Most satellites can move around a little bit to avoid debris or refine their orbits, but what Portal Space is proposing is the equivalent to delivering a new rocket to you to do what’s called a plane change; something you just can’t do today. Maneuvering-wise, a shoe-horned analogy is that you are going from a train, that can control a few factors like speed, but not much else, compared to a plane, where you can control almost anything in the air domain: up, down, fast slow.

Whether Portal solves this problem is unclear. Not much is known of the still theoretical technology being developed. But the idea of removing this maneuverability barrier to the space industry is exciting. More presciently, it means that repositioning satellites as needed becomes more available. Instead of launching multiple satellites to cover all the potential points you may need them, you can launch one and have it maneuver where you need it, when you need it. You save development, operations and sustainment costs. That’s for services on Earth. For in space, you now can start to design systems to clear out debris that execute over several objects instead of having to launch a satellite per object to bring down. It also means that your energy equation to get to where you want to go is also changing, which is very interesting especially in the deep-space exploration domain.

The challenges? The most obvious one is technological, which we won’t speculate on. Unfortunately, the innovations and successes of the current space industry is built on top of the graveyard of tested technologies for space applications (DC-X, Venture Star etc.) But for the sake of the news received, we’ll apply the magic wand and assume Portal delivers something that will work.

One more critical concern will come down to the infrastructure underpinning space coordination, commonly known as Space Situational Awareness (SSA). The concerns of the Kessler Syndrome continues to title reports of various levels sensationalism (partially warranted), and while the infrastructure to track objects is modernizing, when looking closer, those efforts are more accelerated in some areas (such as software interface) than other (such as hardware).

Portal’s innovation means we’re facing the prospects of having to track a lot more satellites that go where they want when the operator wants, without a major cost barrier. How do we make sure they don’t crash into each other? Our ability to track objects and keep everyone informed about where what will be is going to be severely challenged.

Today, much of the SSA infrastructure is built on predictive analysis for the vast majority of satellites: orbital mechanics assume most satellites are non-maneuvering (because it’s simply so expensive to move) and their future orbital position could be predicted with some level of accuracy (space weather models should improve them even more). There are aspects modernizing the more accurate tracking of some suspicious satellites (think spy satellites spying on other satellites), but there still very few actors that play in this field, especially outside of GEO.

A propulsion system that allows you to break from a predictable trajectory means we’re going to need to develop a unified, space control system. A not-very accurate but still useful analogy is simply air traffic control for space; knowing where everyone is and where they plan to go.

And technology is not the only element to challenge this paradigm of ensuring space safety. In the context of the geopolitical challenges, which nation with the resources necessary to make it happen and lead is unclear. Today, for most of the western-world, the American Space Surveillance Network remains the keystone and importantly, free keystone, in SSA.

Fortunately, this complex challenge is a great opportunity for further developments and innovations into SSA, space policy, and other technologies.

Portal Space’s potential technology, if viable, may revolutionize our use of space, much in the way reusable rockets are changing access to space. With low-cost capabilities to move in space means that the space economy of the future will have two of the three major legs needed for the logistics infrastructure (the last remains reentry, which saw a major setback with the Space Shuttle’s retirement). There are use cases to the technology, but that’s true of many ideas; use-cases don’t always translate to financial inflows that fund the development. Portal’s biggest issue is the same as with many other startups: if they build it, will others come? Or will a funding crunch bring down the dreams of a near future mobile space world.

II. Quote of the Week

“The biggest losers in this scenario are the small and medium-sized suppliers that form the backbone of the space industry.”

James Gellert, executive chair at RapidRatings, in response to the tariffs announced by the executive branch of the US government.

III. Immediate Awareness

1 Japanese automaker Honda plans to test a renewable energy technology aboard the ISS, aiming to generate electricity and oxygen using solar energy and lunar soil for future use on the Moon.

2 In a key development for the agency, the ESA will proceed with Phase 2 of its LEO Cargo Return Service initiative without geo-return constraints, allowing contractors greater flexibility in sourcing and design choices across member states.

3 The Portuguese Space Agency has highlighted its €40 million investment in over 65 projects from 2019 to 2024 with dual-use potential, focusing on technologies that serve both civilian and defense sectors, boosting commercial viability for the industry.

4 Taiwan has selected Jiupeng Village in Pingtung County as the site for its national space mission launch center, with plans to promote a nascent strategic space industry enhance technological development by 2030 or 2031.

5 Russia and China are escalating space electronic warfare capabilities by targeting SpaceX's Starlink satellite network, reflecting growing geopolitical tensions and competition over space-based assets.

6 Voyager Technologies announced its acquisition of LEOcloud, a startup focused on space-based cloud computing, aiming to enhance its capabilities in AI-driven analytics, satellite communications, following its pivot into defense.

7 Like almost all other industries, tariffs imposed by the Trump administration and China’s retaliatory actions are disrupting the global space industry, raising costs and straining supply chains, both directly through materials and indirectly, through huge amounts of uncertainty, particularly impacting smaller satellite and spacecraft manufacturers.