In January 2025, the loudest gravitational wave ever recorded hid a second signal nobody knew how to read — and a decade after the first detection, it has just given physicists the spin rate of a black hole's event horizon

The signal arrived at LIGO on January 14, 2025. Two laser beams running down 4-kilometer vacuum tubes in Hanford, Washington and Livingston, Louisiana flickered by a distance roughly ten-thousandths the width of a proton, lasted a fraction of a second, and went still. The disturbance, catalogued as GW250114, turned out to be the loudest gravitational wave ever detected — about three times stronger than the chirp that opened gravitational wave astronomy a decade earlier. Hidden inside it was a signal no one had previously known how to read.

That signal has now been pulled out. In a paper published in Nature on June 24, 2026, Dr. Ling Sun and PhD candidate Neil Lu of the ARC Centre of Excellence for Gravitational Wave Discovery (OzGrav) at the Australian National University, working with Sizheng Ma of the Perimeter Institute, Ornella Piccinni at ANU, Yanbei Chen at Caltech and colleagues in Spain, report the first direct measurement of two fundamental properties of a newly formed black hole’s event horizon: its rotation frequency and its surface gravity.

A louder signal, a sharper window

The collision that produced GW250114 was the same kind that produced GW150914 a decade earlier — two stellar-mass black holes spiraling together at cosmological distance. The primary weighed about 33.6 solar masses, the secondary about 32.2. They fell into each other and left behind a remnant of roughly 62.7 solar masses, spinning at a dimensionless spin of 0.68, with the rest of the missing mass radiated outward as gravitational waves.

What was different was the detectors. Ten years of quantum-precision upgrades have driven instrumental noise down to the point where LIGO can register space-time distortions on scales 10 thousand trillion times smaller than a human hair. That precision is why the same kind of event that once produced a faint chirp now produces a textbook-clean waveform, with a signal-to-noise ratio about three times that of the first detection.

The first detection, GW150914, was announced in February 2016 after a century of speculation and decades of instrument-building. By contrast, the LVK network now routinely observes roughly one black hole merger every three days, and has logged around 300 binary black hole mergers across its four observing runs.

What the direct waves reveal

The new finding hinges on a small piece of the signal that earlier analyses had set aside. “We measured the last sound the black holes made when they crashed,” Lu said in a statement released by ANU. “Hidden within that signal is a small component, called direct waves, that had not previously been well understood. Our new analysis allows us to decipher this component and extract unique information from close to the event horizon.”

The direct wave is gravitational radiation emitted from immediately outside the newly formed horizon, as material spirals across the boundary at the moment a single horizon takes shape from two. Ma, at the Perimeter Institute, proposed searching GW250114 for this signature based on theoretical work showing it should be visible in a loud enough signal. The wave oscillates at approximately twice the rotation frequency of the new horizon, and decays at a rate set by the horizon’s surface gravity.

Those two numbers matter. In general relativity, a non-charged black hole is fully described by its mass and spin. Rotation frequency and surface gravity are the horizon-level expressions of that minimalism — what you would measure if you could hover at the boundary itself. Until now, gravitational wave astronomers inferred them indirectly from the inspiral and ringdown. The direct-wave channel reads them closer to the source.

Sun, the team co-leader, framed the result this way in comments to Space.com: the exceptionally loud signal “can be used as a powerful probe of the remnant black hole’s horizon, allowing us to measure its two fundamental properties: rotation frequency and surface gravity.”

The horizon, briefly

The event horizon is a mathematical surface, not a physical membrane. It emerged from Karl Schwarzschild’s solutions to Einstein’s 1915 field equations, worked out while Schwarzschild was serving with the German army on the Eastern Front during the First World War. He sent the manuscript to Einstein in December 1915 and died a few months later of an autoimmune skin condition contracted in the trenches.

The Schwarzschild radius for the Sun sits at about 2.95 kilometers — roughly 1.83 miles — from its center of mass. For Earth, it is roughly 9 millimeters, about the diameter of a small marble. For the Moon, about 0.1 millimeters. The radius scales linearly with mass; the physics does not change.

A real black hole’s horizon is the surface beyond which no signal can escape to a distant observer. That makes it, in principle, impossible to image with light. Gravitational waves are different. They are emitted by the dynamics of spacetime itself, including the dynamics of the horizon as it forms, rings, and settles.

Black holes come in a vast range of scales. The stellar-mass objects detected by LIGO are tens of solar masses. The ultramassive quasar TON 618, by contrast, has been estimated at roughly 66 billion solar masses based on a 2004 analysis of its Hβ emission line, with more recent work suggesting the figure could be closer to 40 billion. The horizon physics is the same; only the radius changes.

Testing Einstein at the boundary

Loud gravitational wave signals have been used before to probe Stephen Hawking’s 1971 area theorem, which states that the total surface area of black hole horizons cannot decrease. The original test was carried out by Maximiliano Isi and collaborators in 2021 using GW150914 data. A 2025 reanalysis confirmed the theorem at higher confidence.

The direct-wave analysis is a different kind of test. The area theorem asks whether horizons obey a global rule. The direct waves ask what the horizon looks like in detail — its rotation rate, its surface gravity, whether the numbers extracted from the waveform match what general relativity predicts for a rotating black hole described by Kerr geometry. In GW250114, the team reports, the measured values agree with the Kerr prediction.

“These measurements mark a first step towards future tests of general relativity with direct waves,” Lu said. The implication is that as detector sensitivity continues to improve, gravitational wave observatories will be able to look for departures from Kerr geometry — the signature, if it exists, of new physics at the horizon.

There are reasons to look. Some theoretical alternatives to classical black holes predict horizons that behave differently, or no horizon at all. Gravastars replace the singularity and the horizon with an exotic surface. Quantum gravity proposals predict subtle deviations in how a horizon rings down after a merger. Direct waves give experimenters a new channel to constrain those ideas.

From discovery to metrology

Ten years separate GW150914 from GW250114. The first chirp was a confirmation of a 100-year-old prediction; it told observers gravitational waves existed. The current loudest signal is something closer to a measurement instrument — a probe pointed at an object that emits no light at all and resolves features the previous generation of detections could only hint at.

Most events in the catalog will not be loud enough for this kind of horizon analysis. By Lu and Sun’s own description, GW250114 was an exceptional case: a strong, nearby, clean merger that happened to fall well within the band where LIGO is most sensitive. The matched-filter signal-to-noise ratio of the direct wave component alone was around 16 in the LIGO Hanford detector — high enough to isolate, low enough that quieter events will not yield it.

What matters scientifically is that the technique now exists. The next decade of gravitational wave astronomy will be the first in which the loudest detections are routinely loud enough to read at the horizon.

Future detectors will push further. The proposed Cosmic Explorer in the United States and Einstein Telescope in Europe are designed to be roughly an order of magnitude more sensitive than the current network. At that level, very loud signals stop being exceptional, and the direct-wave analysis can be applied across populations of black holes rather than to one merger that happened to be uncommonly close.

For now, the result establishes that the loudest crashes carry information that earlier, quieter crashes did not. The horizon — the surface theorists have argued about since 1916 — has started to give up numbers.

The wave that carried those numbers traveled for roughly a billion years before flickering through the mirrors in Hanford and Livingston. By the time anyone read it, the merger itself was a billion years gone. Somewhere far away, the black hole that formed is still there, still spinning, still dragging spacetime around with it — the surface that researchers can now measure quietly continuing to turn.