AI Spots Wandering Supermassive Black Hole 760M Light-Years Away

For space surveillance and missile warning operators, the ability to detect faint, transient events in a sea of background noise is the holy grail. NASA's newly reported discovery—a star-shredding black hole found not at a galaxy's heart but in its empty outskirts—was only possible because of an AI algorithm that sifted through half a million flashes per night. That same automated anomaly detection capability is increasingly central to military space programs, where recognizing a hostile satellite maneuver or a debris collision in real time can mean the difference between a near miss and a catastrophic strike.

In a significant milestone for survey astronomy, NASA's Neil Gehrels Swift Observatory has captured the complete disruption of a star by a supermassive black hole lurking in the outskirts of a galaxy 760 million light-years away. The event, initially spotted in late 2025 by the Zwicky Transient Facility (ZTF) as a bright flare in the constellation Cetus, was later confirmed as a tidal disruption event (TDE) by Swift's multi-wavelength instruments. This is the first TDE to be unambiguously associated with a wandering supermassive black hole—one that is not anchored at the center of its host galaxy. The black hole, estimated to weigh about one million times the mass of our Sun, was detected only because it happened to tear apart a passing star, creating a luminous accretion disk that could be seen across the cosmos.

“

In a significant milestone for survey astronomy, NASA's Neil Gehrels Swift Observatory has captured the complete disruption of a star by a supermassive black hole lurking in the outskirts of a galaxy 760 million light-years away.

The discovery, published on July 27, 2026, in The Astrophysical Journal Letters, underscores a paradigm shift in how astronomers hunt for otherwise invisible black holes. Traditionally, supermassive black holes are identified by their gravitational influence on nearby stars and gas, but off-center black holes have been nearly impossible to find. TDEs offer a fleeting beacon, yet their rarity and unpredictable locations make them exceptionally hard to catch. The breakthrough came through the use of an artificial intelligence tool called tdescore, which has been systematically processing the half-million transient alerts ZTF generates every night since 2024. The algorithm automatically flagged this event despite its unusual position far from any galactic core—a location that would typically cause human classifiers or even rule-based filters to dismiss it as noise or a supernova.

For the space and defense communities, this success illustrates the maturation of AI-driven anomaly detection in large-scale surveillance networks. The ability to sift through massive datasets in real time, identify relevant events, and trigger follow-up observations is directly transferable to space situational awareness (SSA) and missile warning systems. The astrophysical event itself acts as a field test for algorithms that could one day monitor satellite maneuvers, space debris collisions, or hostile activities in orbit. Furthermore, the collaboration between ground-based optical surveys (ZTF) and space-based multi-wavelength observatories (Swift) mirrors the layered sensor architectures being deployed for domain awareness in cislunar space and beyond.