On July 2, 2025, space telescopes monitoring the sky for brief, one-and-done flashes of high-energy light saw something that nobody expected: a gamma-ray burst (GRB) that came back again and again, stretching what is usually a single “burst” lasting seconds to minutes into an all-day event. NASA’s Fermi spacecraft triggered on multiple gamma-ray episodes from the same patch of sky over several hours, and other satellites soon reported compatible detections. Compared to the known population of GRBs that have been studied for decades, this was an outlier beast of a different species.

At first, the event’s location near the crowded plane of the Milky Way made it tempting to suspect something closer to home, located in our own Galaxy. But follow-up imaging overturned that assumption. Observations with the Very Large Telescope (VLT) in Chile narrowed down the position and, together with Hubble and JWST, revealed that the transient was coincident with a dusty, irregular host galaxy. The distance is extreme: the light from the explosion began its journey roughly 8 billion years ago. In other words, whatever happened was not a local flare—it was a truly cosmic-scale detonation, or, rather, a string of detonations.

The duration of this event was not the only weird thing about it. Archival data showed that low-energy X-rays were already present almost a day before the main gamma-ray fireworks—an “X-ray precursor” that is hard to reconcile with standard models of GRBs. Meanwhile, the gamma-ray behavior itself looked like a stuttering engine. Fermi detected a sequence of short flares separated by long gaps, collectively implying multi-hour activity from a central engine rather than the single, clean explosion typical of such events.

So, what could power an event that (1) repeats, (2) lasts for hours to a day, and (3) shows X-rays both before and after the gamma-ray fireworks? Two families of ideas have dominated the discussion. One idea keeps it in the GRB family but pushes the engine to extremes. Typical GRBs arise from the death, or collapse, of massive stars, which can produce a narrow, relativistic jet that emits gamma rays. Perhaps some aspect of the collapse, either the stellar type or the nature of the compact remnant(s) left behind could produce a central power source that simply refuses to shut off on normal timescales. The other main idea is an event completely unlike traditional GRBs and instead invokes a star wandering too close to a black hole, being torn apart, and feeding a jet aimed toward Earth. Such phenomena, known as tidal disruption events, were first predicted in the mid-1970s, but only detected twenty-five years later. Currently, we find a handful of these energetic shredding events each month, but what would make this tidal disruption event so different from the previously observed examples? The catch is that each scenario explains part of the puzzle and strains against the rest, leaving GRB 250702B as a genuine classification stress-test for high-energy astrophysics.

NuSTAR catches the engine in the act, days later

NuSTAR is built to detect high-energy X-rays, and in this event, it provided an important piece of forensic evidence. The system stayed restless well after the headline gamma-ray activity. A comprehensive X-ray campaign led by Brendan O’Connor (Instituto de Radioastronomía y Astrofísica, Mexico) combining data from the NuSTAR, Swift, and Chandra satellites found that the X-ray emission faded steeply overall. But, crucially, Swift and NuSTAR continued to detect rapid X-ray flares out to about two days after discovery. That short-timescale variability is difficult to attribute to a simple, smoothly decaying afterglow alone; instead, it points to ongoing, intermittent activity from the central engine long after standard GRB models would expect the fireworks to be over.

NuSTAR’s high-energy X-ray spectrum also helped connect competing interpretations to actual physical constraints. In the analysis jointly led by Gor Oganesyan and Annarita Ierardi (GSSI, Italy), and Elias Kammoun (Caltech, USA), the Swift lower-energy X-ray decline is shown to be extremely rapid over the first days but also shows persistent flaring activity. The NuSTAR high-energy X-ray measurement (taken about ten days after the trigger) is consistent with that same rapid fade. One idea is that this event could be associated with a "micro-tidal disruption event" in which a star was torn apart by a stellar-mass black hole, i.e., a black hole with a mass approximately ten times that of the Sun, rather than traditional tidal disruption events that involve black holes with masses thousands to millions of times that of the Sun. In short, NuSTAR did not just add data to the pot—it anchored the high-energy X-ray behavior that makes the event so hard to explain simply as a standard GRB or a standard tidal disruption event.

Where things stand now is both satisfying and unsettling. GRB 250702B is almost certainly extragalactic, almost certainly powered by a jet, and almost certainly driven by an engine that stays active far longer than a canonical GRB. But whether that engine was a star being shredded by a black hole or an unprecedented variant of a GRB progenitor remains an unanswered question, precisely because the observations pull in both directions. Resolving the origin may require what high-energy astronomers love most: the next strange event, caught early, followed deeply, and watched closely until the power source finally, and unambiguously, goes dark.

Author: Elias Kammoun (Postdoctoral Researcher, Caltech)

Full animation on NASA SVS: https://svs.gsfc.nasa.gov/14916