Astranis wants to watch every satellite in the GEO belt. The interesting part isn't the spying — it's what one watcher costs.

The satellite Astranis unveiled this week is built to sidle up to other satellites, photograph them from several angles, and move on. It is called Perceptor, it will live in geostationary orbit, and the company's pitch for it comes with a line worthy of a thriller. "There are blind spots that will be used against us if not eliminated," said chief executive John Gedmark. It is a good line, and it is the part of the announcement to be most suspicious of, because the interesting thing about Perceptor is not that it can spy. Governments have flown satellites to inspect other satellites for years. The interesting thing is what one of these costs, and where it comes from.

Space-domain awareness — the military's term for keeping track of what is where in orbit and what it is doing — has until now been an expensive, bespoke business. Perceptor is an attempt to make it a catalogue item. That is a business story dressed as a security one, and the business story is the one that will decide whether any of this flies.

Start with the invoice

The reason to pay attention is not the sensor. It is the bus — the satellite underneath the sensor. Perceptor is built on Astranis's microGEO platform, the same roughly 350-kilogram spacecraft the company already builds to sell broadband from geostationary orbit. A microGEO satellite weighs about a twentieth of a traditional geostationary bird and, by the company's account, costs on the order of a tenth as much. A conventional GEO satellite runs to the high hundreds of millions of dollars and takes three to five years to build. Astranis builds its little ones in twelve to eighteen months. Nobody has published a firm sticker price for a Perceptor, so treat the exact figure as unknown, but the platform it rides on is a satellite you can measure in tens of millions, not hundreds.

Hold that number against the thing it is meant to replace. The United States already flies inspector satellites in geostationary orbit under a program called GSSAP — the Geosynchronous Space Situational Awareness Program — a small number of exquisite, purpose-built spacecraft that quietly maneuver up close to other countries' satellites and take a look. They work. They are also few, expensive, and built the old way. GSSAP is the bespoke suit. Perceptor is Astranis proposing to sell space-domain awareness off the rack.

Why cheap changes the math, not just the price

Here is the shift the renderings tend to hide. The hard part of watching geostationary orbit was never building one good camera. It was covering the ring. Geostationary orbit is a band roughly 35,800 kilometers up, circling the whole planet, and the objects worth watching are spread all the way around it. A single inspector, however capable, can only be in one arc of that ring at a time. To keep eyes on a target — and to look at it from more than one angle, which is what tells you whether that dish is pointed at you — you need to reposition, and repositioning across the belt burns propellant and, with it, the satellite's life.

That is a quantity problem, and quantity is exactly what a bespoke, hundreds-of-millions-of-dollars satellite cannot give you. You do not build a fleet of those. You build two or three and ration them. Drop the per-unit cost by an order of magnitude and the arithmetic changes character: now you can afford to put several watchers around the belt, let each one spend its maneuvering budget freely, and treat any single spacecraft as expendable enough to send somewhere risky. Astranis says multiple Perceptors can observe the same spacecraft from different positions at once, which complicates anyone's attempt to duck the gaze. That capability is not a sensor breakthrough. It is a consequence of the price.

The hard part was never building one good camera. It was covering the ring — and covering the ring is an arithmetic problem that only gets solved by making each watcher cheap enough to lose.

Follow the money, and it leads to a contract vehicle

So who pays? Not, in any near term, a commercial market. The buyer for eyes in geostationary orbit is the government, and the relevant line item already exists. In April the Space Force named fourteen companies — Astranis among them, alongside Anduril, Lockheed Martin, Northrop Grumman, L3Harris, Boeing's Millennium, Sierra Space, True Anomaly and others — to compete for orders under a program called Andromeda, formerly RG-XX. It is the planned replacement for GSSAP: instead of a few large custom satellites, a constellation of smaller, more mobile, lower-cost commercial ones, some able to refuel and stay on station longer. The contract is structured as an indefinite-delivery vehicle running up to ten years, and its ceiling has already been raised from an initial $1.8 billion to reportedly north of $6 billion.

That is the appropriations story underneath Gedmark's blind-spots line, and it is worth being precise about what it is. A contract ceiling is not a check. It is permission to spend up to a number, parceled out later as individual task orders that companies still have to win. The $6 billion is the size of the door, not the amount that has walked through it. But the door is what tells you the business case is real: the Pentagon has decided, on the record, that it wants space-domain awareness delivered as many cheap commercial satellites rather than a few exquisite ones, and it has set aside the machinery to buy them that way. Astranis has spent this year raising money for exactly that pivot — a $450 million round earlier in 2026 aimed partly at manufacturing capacity and government procurement. Perceptor is the product that turns that permission into revenue.

The line to be suspicious of

Return to the thriller line, because it is doing a job. "Blind spots that will be used against us" is threat marketing — the space-business equivalent of inspiration, which is always a red flag in a business case. It is not that the concern is fake; geostationary orbit really is getting more crowded and more contested, and other countries really do fly satellites that shadow ours. It is that fear is the least reliable part of the pitch, because it is the part that does not have to be true to sell. The sturdier sentence in the announcement is the boring one: Gedmark's promise of a "proven, on-orbit platform and ability to produce at scale." Translated, that is a company saying it already has a working factory and would like a government to help keep it busy. That is a better reason to believe in Perceptor than any blind spot.

The honest version of the skepticism is not that the security case is invented. It is that Astranis is a broadband company that has found a second, government-funded use for the satellite it already knows how to build, at a moment when the broadband business is hard and the defense budget is generous. There is nothing wrong with that. It is how a lot of good space companies survive. But it means the thing to watch is the factory and the task orders, not the render.

Where the math might not close

A few costs the low sticker price does not capture, and any of them can bend the business case:

• Cheap to build is not cheap to operate. Space-domain awareness is a service, not a satellite. Repositioning a watcher across the belt spends electric propulsion and shortens its life, and the imagery it collects is worthless until someone processes it into a decision. The recurring cost — propellant, ground stations, analysts — is the invoice the launch photo never shows.
• An IDIQ ceiling is not backlog. Fourteen companies are competing for those task orders. Being named to the vehicle is a ticket to bid, not a booked sale, and Astranis is bidding against primes with far deeper defense relationships.
• A watcher is also a target. A satellite built to maneuver up close to someone else's spacecraft is, to that someone else, a provocation — the same behavior that makes GSSAP useful is the behavior other countries call hostile. Cheap, numerous inspectors make the geostationary belt more crowded with things designed to shadow each other, and that is a stability cost that lands on everyone, not a line on Astranis's balance sheet.
• 'Produce at scale' only pays at scale. The whole cost advantage rests on volume. If the orders come in ones and twos, a mass-production bus loses the economics that justified it, and Perceptor becomes an expensive way to do what GSSAP already does.

None of that sinks the idea. The core bet — that watching orbit is an arithmetic problem best solved by making each watcher cheap enough to lose — is a sound one, and it is the same bet, pointed at defense, that made Astranis's broadband satellites interesting in the first place. The question is the one that decides every space business: does the math close at the volume the customer actually buys? Astranis has built the cheap satellite. Whether space-domain awareness becomes a catalogue item now depends on how many the Space Force orders, and at what price per copy. Watch the task orders. The blind spots will take care of the marketing.