“Small satellite” is a definition built on mass, not ambition. The general cutoff is 500 kilograms, and below that the categories keep subdividing into minisatellites, microsatellites, nanosatellites, and CubeSats measured in units the size of a loaf of bread. What none of those categories tell you is what the thing can actually do, which is the whole point. If it can be put on a small platform and launched… it will be.
I’m writing this from the Salt Palace in Salt Lake City, where roughly 5,000 people have gathered for the 40th Annual Small Satellite Conference. This year’s theme is “The Clouds Above the Clouds,” and the technical focus is squarely on constellations – groupings of satellites that work together to provide more data collection than single satellites can do alone, enabling better temporal resolution, more opportunities for calibration, and solving for tricky tasking. Forty years is long enough for a field to stop being an experiment and start being an industry, and walking the exhibit hall, you can feel both versions of it in the same room.
I wrote this down yesterday, in between sessions, and I stand by it:
The industry lives in that blurry space between defense and science, between academia and private industry, and between staggering profits and shoestring budgets.
The numbers back up the blurriness. Government and defense still account for roughly 54 percent of small satellite end use in 2026, so the checks are still largely being signed by agencies. But the Earth observation smallsat segment is projected to grow from about $2.14 billion in 2026 to $6.90 billion by 2034, and commercial end users are growing faster than any other buyer segment at about 17 percent annually. Earth observation is now the fastest-growing application in the entire small satellite market, outpacing communications, driven by agriculture, insurance, and logistics customers who want imagery several times a day instead of once a week.
But in the in-between, there’s the answer to some key questions. What if I could see that are better? What if I had the ability to try something out? What new innovations might reveal something we’ve never seen? This is where small satellites fit, at least for now.
For most of the last twenty years, the division of labor was clean. If you needed detail, you flew. If you needed scale, you bought imagery. Aerial programs delivered sub-10 centimeter ground sample distance, oblique views, and controlled acquisition conditions. Satellites gave you 30 to 50 centimeters at best and you took whatever the overpass gave you.
The optical side is closing the gap slowly and expensively, but the real technology to watch is SAR (synthetic aperture radar). Radar is where smallsats have actually changed what’s available. ICEYE-X1 was the first SAR satellite to fly under 100 kilograms, and Umbra’s satellites are 70 kilogram microsatellites sized to fit an ESPA ring. Both are now delivering spotlight imagery down to roughly 16 centimeters, from platforms that weigh less than the person operating them.
But resolution is the least interesting part of this, honestly. The more useful question is not “is the pixel small enough” but “can the aircraft even get there?” For situations where the answer is no, satellites might soon be a realistic option.
Imagine after a hurricane, a wildfire, a chemical release, or a levee breach – often the airspace over the thing you need to see is the airspace you cannot fly. Or instances where it’s literally too smoky, foggy or dark. Optical aerial acquisition needs to wait for a clear day and reasonable sun angle, SAR satellites don’t. Flying a crew and an aircraft to a remote coastline or an island can sometimes cost more in logistics than in flight hours. In hydrography, satellite-derived bathymetry has been deployed in difficult-to-navigate or remote territories where data gaps persist or no environmental footprint is permitted, delivering wide shallow-water coverage without mobilizing equipment or personnel. Much of that work has run on large government missions, and the smallsat contribution is less about a new way to measure depth than about how often you can look.
The time scale matters, too. Wide-area change detection has historically been constrained by how often you could afford to refly. When the ability to revisit becomes daily and delivery becomes hourly, change detection stops being a project and starts being a monitoring service. That is a different business model, and it is one that geospatial services firms are well positioned to sell.
So what do geospatial professionals really need to know?
Start with the framing. The question is not whether satellites are going to replace aerial imagery or drone flights. That framing has never been useful and it isn’t useful here. The better question is which layer of the workflow this occupies. In hydrography, satellite-derived bathymetry isn’t competing with bathymetric lidar, it’s scoping where the lidar should fly.
Second, notice what buyers are asking about. Procurement conversations in this market are increasingly built around data latency and API access rather than resolution alone. It’s the same conversation our field has been having about reality capture pipelines, cloud-hosted point clouds, and who owns the workflow between the sensor and the decision.
Third is that we need to keep asking what is flying right now versus what is on a roadmap. A funding round is not a constellation and a single commissioned satellite on a pilot project is not going to give you daily data. The gap between the pitch and the reality is where most disappointment in this industry has historically come from.
Which brings me to where I think the opportunity actually sits. This industry is extremely good at getting sensors into orbit and getting better every year at getting data down quickly. What it is less good at is the thing our profession has spent more than a century on – actually assessing that accuracy against claims. Somebody has to be able to tell when data is good enough to run a monitoring program but not good enough to set a boundary. Somebody has to validate the model, check the control, and write the caveat. The satellite companies are not going to do that work, the users will.
I wrote that small satellites live in the “in-between” and they still do, even 40 years in. While we’re not to the point where most of us need to examine this branch of tech, it’s worth keeping an eye on. There’s always plenty of room for people to ask whether the measurement holds up – and that’s something our industry has never been shy about asking.
