Most reconnaissance satellites sit in one orbit and photograph the ground. The two the Pentagon just ordered are designed to move, and to photograph each other's neighbours.

On 18 September the Defense Innovation Unit announced that Northrop Grumman and True Anomaly will build prototype satellites for GHOST-R — GEO High-Resolution Optical Space-Based Tactical Reconnaissance. The job is to manoeuvre around geosynchronous orbit and take detailed pictures of other spacecraft, friendly and otherwise.

The prototypes are due to launch in 2028. If they work, they become the operational RG-XX fleet from 2029.

Why Satellites Now Need to Manoeuvre

Geosynchronous orbit sits about 36,000 kilometres up, and it is where the expensive things live: communications relays, early-warning satellites, national technical means. It is also crowded and increasingly contested.

The US already watches it, using the Geosynchronous Space Situational Awareness Program — GSSAP — a small fleet of classified inspector satellites built by Northrop Grumman. GSSAP works. What it is not is cheap, numerous, or easy to share with allies.

Since 2024 the Space Force has been looking for a lower-cost commercial alternative to augment it, partly because US Space Command wanted more unclassified opportunities to work with allies and industry, and partly because it wanted two specific things: satellites that can move, and better awareness of what is actually up there.

The Two Approaches

Northrop Grumman is the incumbent. It built GSSAP, and its GHOST-R design leans on an existing high-rate production bus called ESPASat-L, with optical payloads supplied by TRL 11, a company based in Irvine, California. The pitch is essentially: we have done this before, and now we can do it at volume and at a lower price.

True Anomaly is the challenger, and its case rests on a spacecraft called Jackal that has already been flown. During the Space Force's Victus Haze responsive-space mission, Jackal approached a target satellite built by Rocket Lab, imaged it, characterised it, and — in the company's phrasing — engaged in a space-to-space pursuit of the vehicle.

True Anomaly's own film on the Jackal spacecraft.

What the Prototypes Must Prove

The demonstrations are specific. Each satellite must show it can collect high-resolution imagery and conduct weekly observations of target satellites. In parallel, both companies must submit a plan for scaling their design into a full constellation capable of tracking every publicly catalogued object in geosynchronous orbit.

That last requirement is the tell. This is not a science project; it is a pilot for an industrial buy.

The Money Behind It

GHOST-R feeds RG-XX, and RG-XX sits inside something much larger. In April the Space Force selected 14 companies for an industry consortium called Andromeda, where they compete for task orders worth up to $6.2 billion. That covers both the manoeuvrable RG-XX reconnaissance satellites and a separate surveillance fleet designated SG-XX.

Alongside the hardware, the Defense Innovation Unit has posted a solicitation for software able to ingest and fuse large multi-source data streams — satellite imagery, geospatial data, live video — and automatically distinguish friendly objects from threats. Building satellites that can look closely at things is one problem. Working out which of the things deserve looking at is the other.

For aviation readers there is a familiar shape to all of this. A small number of exquisite, expensive, classified platforms is being supplemented by a larger number of cheaper, commercially derived ones that are expected to be attritable and quickly replaced. The same argument is currently reshaping combat aircraft. It turns out orbital mechanics does not exempt you from procurement fashion.

Sources: Air & Space Forces Magazine reporting by Courtney Albon, with quotes from Col. Bryon McClain, Gen. Stephen Whiting and Even Rogers; Defense Innovation Unit announcement; Northrop Grumman and True Anomaly statements.