Where Missile Sensors Learn to Watch Satellites
The imaging hardware now circling Earth began its working life pointed at targets on the ground, not at spacecraft gliding overhead. Anduril Industries built its long-wave infrared sensor by adapting the imaging technology it already uses in autonomous weapons, then flew it on a small satellite to test whether the same eyes could spot satellites in the dark. The company funded the effort with its own capital rather than a government contract, a detail that matters for how the capability will reach the market.
The satellite, named Anduril-216, launched in July 2026 on a SpaceX rideshare mission. [1] It carries three technology demonstration payloads: an edge-compute processor, a power-management system built to adapt across different mission types, and the long-wave infrared space sensor. Anduril reported the mission’s success on Tuesday. The spacecraft flies at an altitude of 590 km and spends about one-third of each orbit in Earth’s shadow. [1] That shadow is the whole point. Optical tracking has a structural gap: when a satellite passes through eclipse, it is no longer lit by the Sun, and ground-based visible sensors lose it. A spacecraft that maneuvers during that darkness can slip its watchers. Gokul Subramanian, Anduril’s senior vice president of engineering, described what the new sensor changes: “This is a unique sensor that allows us to take imagery in long-wave [infrared]. What’s unique about that is it’s got a huge dynamic range. It has very low solar exclusion, so we’re blocked by the Sun very rarely, and we can see things in full dark.” [1] The sensor works like a camera that reads body heat instead of needing a floodlight. The sensor detects heat signatures from spacecraft even during eclipse, which Adam Thurn, vice president and general manager of orbital systems at Anduril, framed as closing a specific gap: “Right now, we don’t have ground sensors in our adversaries’ sovereign lands, and then in the eclipse period where we are relying heavily on visible sensors, the satellites that are not illuminated by the Sun, we can’t see those. With the long-wave infrared space sensor, we’re looking to close that gap.” The capability offers what he called a way “to witness when an adversary or other resident space objects are maneuvering during those eclipse periods, where in the past we’ve been fairly blind.”
The Sensor That Thinks Before It Downlinks
The tracking problem has two halves, and Anduril’s satellite carries hardware for each. The first is seeing; the second is deciding what was seen. The edge-compute module on Anduril-216 is designed to fuse data from different sensor types — infrared and visible imagers among them — and generate tracking solutions on board, without waiting for ground analysis or intervention. Instead of shipping raw pixels to analysts, the spacecraft does part of the interpretation itself.
Anduril released a single image from the sensor showing an infrared view of the Moon above Earth’s horizon. The Moon image, Thurn said, offers just a taste of what the instrument can do. He then sketched the instrument’s range: “The long-wave infrared capability we have on orbit can do both unresolved and resolved imagery.” That opens a further path. If the sensor flies close enough to another satellite, it could be used to inspect it.
The current focus is narrower than that eventual use. Thurn said Anduril is optimizing the tracking sensor for low-Earth orbit, while noting it could serve other orbits and eventually cislunar space near the Moon. The company’s near-term business model is narrower still: Subramanian said Anduril’s focus is on acquiring space domain awareness data for sale to the US military, not on selling entire satellites or sensors to the government. “At present, Anduril is not offering this system as a merchant supplier,” he said. “That could change in the future.”
A Data Business Built on Borrowed Platforms
The spacecraft platform itself was not built by Anduril. Apex Space, a satellite-manufacturing startup, supplied the bus for Anduril-216. Subramanian said Anduril has no plans to start its own satellite production line. Instead, the company intends to buy third-party platforms and integrate them with its sensor and compute modules for a range of missions. Someone else builds the chassis, Anduril supplies the eyes and the judgment.
The mission design followed a similar logic of starting small. “You’re seeing that today with Anduril-216, where we started with a simple mission, which was how do we track satellites, especially in these eclipse periods where the US currently does not have the ability to maintain custody,” Subramanian said. “And we designed a mission around that.” Custody — the continuous knowledge of where a spacecraft is and what it is doing — is the product being tested, not a finished service.
The military’s interest runs along two distinct strands. The Space Force tracks satellites for two reasons: commanders want to know whether an adversary is moving a spacecraft in a threatening manner, and global operators need collision alerts when orbital paths cross. Anduril’s demonstration targets the first strand and brushes the second.