On March 4, 2024, a radio telescope in South Africa caught something that lasted roughly a millisecond. That was the end of the easy part. The flash, named FRB 20240304B, looked so distant that nothing on Earth could see the galaxy it came from. So astronomers did what you do when the universe hides something: they called in the James Webb Space Telescope.

Webb found it. Its infrared camera spotted a faint galaxy almost exactly at the burst's position, and its spectrograph measured the telltale signature of distance, a cosmological redshift of 2.148, plus or minus 0.001. That number is the record. The burst happened when the universe was only about 3 billion years old, roughly a quarter of its current age, and its signal traveled for more than 10 billion years to reach us. It more than doubles the previous distance record for a localized fast radio burst.

The findings, led by Manisha Caleb of the University of Sydney and published in Science on October 8, do more than set a record. The galaxy Webb found was not the galaxy anyone expected, and that mismatch has consequences for one of astronomy's favorite unsolved mysteries: what makes these flashes at all.

An 18-month hunt for a galaxy nobody could see

Fast radio bursts were discovered in 2007, and the field has never lacked detections. What it lacks is distance. Radio telescopes like MeerKAT can localize a burst to a patch of sky, but converting that location into a distance means finding the host galaxy and measuring its redshift, and the farther away the burst, the fainter the host.

This one was invisible from the ground. The largest optical telescopes on Earth saw nothing at the position. That by itself was a clue: whatever made this flash was very far away. Webb's NIRCam finally detected a faint galaxy about 0.3 arcseconds from the burst position, with a 97.5 percent probability of being the true host, and NIRSpec delivered the redshift.

The pairing is the point. MeerTRAP on MeerKAT discovers and localizes distant bursts; Webb studies the faint hosts ground-based telescopes cannot touch. Together, they can reach the young universe.

The wrong galaxy

Enjoying this story?

Get the five most important stories in tech, every morning. Free.

Most known FRB hosts are big, mature, star-forming galaxies. That is the template the field has worked with, and it fed the expectation for this burst too. Instead, Webb found something the team described as roughly 1,000 times less massive than expected: a small, clumpy dwarf galaxy with a stellar mass of about 10 million suns, a star-formation rate of about 0.2 solar masses per year, and a metallicity only about 10 to 20 percent of the sun's.

Weighing in with what was, in galactic terms, almost nothing: that is the sentence the whole paper pivots on. "We thought it would be a big, nicely formed galaxy with lots of stars, and instead it was a little dwarf galaxy, although it was actively forming stars," Caleb said. And the dwarf was young in the most extreme sense. Its rate of star formation suggested that the majority of its stars may have formed within just 30 million years, a blink of cosmic time.

We thought it would be a big, nicely formed galaxy with lots of stars, and instead it was a little dwarf galaxy, although it was actively forming stars.

The timing matters. Redshift 2.148 lands this galaxy squarely at cosmic noon, the period in the history of the universe when star formation was at its peak. Astronomy has FRBs at many distances, but this is the first time the phenomenon has been pinned down to that era. It means whatever makes FRBs was already at work when the universe was young, and it was happening in galaxies very unlike the ones astronomers had been looking at.

The merger theory takes a hit

FRB 20240304B: Chasing a Flash Across Deep Time

How astronomers traced a millisecond-long radio flash to the young universe.

Mar 4, 2024Detection
MeerKAT's MeerTRAP project catches a millisecond-long radio flash at 900 to 1,700 megahertz.
2024-2026The hunt
The host galaxy is too faint for ground-based telescopes. Astronomers aim Webb at the position.
Oct 2026Host found
NIRCam spots a faint dwarf galaxy; NIRSpec measures a redshift of 2.148, plus or minus 0.001.
Oct 8, 2026Published
The results appear in Science: the most distant fast radio burst ever localized, more than doubling the old record.

Redshift 2.148 means the light left its galaxy about 10.4 billion years ago, when the universe was roughly 3 billion years old.

Here is where the wrong galaxy becomes the most important part of the story. One leading theory holds that fast radio bursts come from the merger of two neutron stars. The physics is elegant: two dead stars spiral together over billions of years, collide, and release an enormous burst of energy. But the timeline is the problem. Binary neutron stars typically need at least a billion years to merge.

This burst came from a galaxy where most of the stars are less than 30 million years old. As Caleb told reporters, the team's analysis suggests it is very unlikely that this FRB was produced by a merger. There simply had not been enough time.

The alternative looks better in this light. The other leading theory says some FRBs come from young, highly magnetized neutron stars called magnetars. When a massive star explodes as a supernova and leaves behind a magnetar, violent events like starquakes can fire off a burst relatively soon afterward, with no long delay required. A young, intensely star-forming dwarf galaxy is exactly the kind of place where a newborn magnetar should exist.

One caution, emphasized by the researchers: this does not settle the origin of all fast radio bursts. Repeating bursts have been convincingly tied to magnetars before, but one-off bursts may come from several channels. What this result does is rule out the merger channel for this event and show that the FRB population at cosmic noon lives in very different environments than the nearby population astronomers have studied most.

A flashlight for the cosmic web

A telescope facility under the night sky
The flash lasted a millisecond and traveled for more than 10 billion years before MeerKAT caught it. Finding where it came from took Webb's infrared eyes. (Photo: Radio Astronomy Journal)

The distance record was not the only payoff. A fast radio burst is, as co-author J. Xavier Prochaska of UC Santa Cruz put it, almost like a cosmic flashlight: it lights up everything along its path and carries an imprint of every bit of matter it travels through. This burst crossed more than 10 billion years of intergalactic space, and the team could read two of the structures it passed in its signal.

One was the nearby Virgo Cluster, only about 54 million light-years from Earth. The other was a previously unknown galaxy cluster at a redshift of 0.3, about 3.5 billion light-years away, discovered effectively by accident because this ancient flash happened to shine through it. This is the method astronomers have dreamed about: using the most distant FRBs as probes of the cosmic web, the otherwise invisible scaffolding of matter between galaxies.

Each distant burst samples everything between us and the young universe. As more of them get localized with Webb-class follow-up, astronomers can start measuring the distribution of invisible matter on the largest scales, turning a mystery into a measuring instrument.

What comes next

The real significance of FRB 20240304B may be that it is a proof of method. "What is particularly exciting about our result is that we've now demonstrated that we can identify and study an FRB from when the universe was young," Caleb said. If MeerTRAP and its successors can find these distant flashes routinely, and Webb (and eventually its successors) can nail down their hosts, astronomers could build a census of FRBs across cosmic time.

That census would answer the questions this discovery only sharpens. Are bursts at cosmic noon systematically different from nearby ones? Does the merger channel dominate later while magnetars dominate early? How much invisible matter can these flashes illuminate?

For now, one flash, one millisecond, has done a remarkable amount of work. It set a distance record, disproved a theory for one event, discovered a galaxy cluster by accident, and gave astronomers their first look at the FRB era of the young universe. Not bad for a signal that was over before anyone noticed it had begun.