In April 2023, the LIGO, Virgo, and KAGRA detectors caught a gravitational wave from the merger of two black holes. With masses of 137 and 103 solar masses, GW231123 became the most massive event ever observed. Both holes fall in the 'pair-instability gap', a range where ordinary stellar evolution should not leave black holes. The component spins, 0.90 and 0.80, are also anomalously high for such giants. Where did this pair come from?

Astrophysicists have long considered primordial black holes, objects born in the first instants after the Big Bang from density fluctuations. Such holes can grow by accretion, that is, by gas falling onto them from the surrounding environment. The new work shows that starting from two modest seeds of 85.5 and 59.1 solar masses, accretion grows them to the observed 137 and 103 solar masses. Meanwhile, the infalling gas spins up the holes, raising their spins to 0.90 and 0.80, exactly as in the LIGO-Virgo-KAGRA data.

The model is not just a fit to the parameters. It predicts the merger rate: to explain one GW231123, the fraction of primordial black holes in dark matter must be about $10^{-3}$. This value sits right at the edge of regions excluded by X-ray and cosmic microwave background observations, so the hypothesis is not yet contradicted by data, but it is not confirmed either. The upcoming O5 run should catch about 20 similar events if the interpretation is correct.

The primordial black hole scenario explains both masses and spins without exotic astrophysical channels like hierarchical mergers in dense clusters. If it is correct, primordial black holes could make up a significant fraction of dark matter in the tens-of-solar-masses range. The Einstein Telescope will be able to see mergers at redshifts z > 15, where stellar black holes did not yet exist. That would be the decisive test.

Modest seeds growing into giants under a rain of gas, that is one way to describe the birth of GW231123. The next gravitational waves will show how close this picture is to reality.