We just heard a black hole's surface ring

TL;DR

• A team led by Dr. Ling Sun and PhD candidate Neil Lu (ANU / OzGrav, Australia) with collaborators at the Perimeter Institute (Canada), Caltech (US) and Spanish gravitational-wave groups have published in Nature the first direct measurement of two defining properties of a black hole's event horizon: its rotation frequency (Ω_H) and surface gravity (κ).
• The result comes from GW250114, recorded by LIGO on 14 January 2025 — the loudest, cleanest gravitational-wave signal ever detected, with a signal-to-noise ratio roughly three times that of the first 2015 detection.
• Hidden inside the post-merger "ringdown" was a previously-unrecognised component — a "direct wave" — predicted by recent theory but never before isolated from data.
• The measured horizon properties match the predictions of the Kerr black hole solution from general relativity to the precision currently available.
• This is the first time anyone has measured an event horizon, rather than inferred its existence. It does not prove black holes are exactly Kerr objects; it sharply narrows the room for them not to be.

A century of inference, ending

Karl Schwarzschild solved Einstein's field equations in early 1916, in a notebook, on the Eastern Front. He died of pemphigus that May. The solution he produced described a non-rotating, spherically symmetric mass and contained a radius — now called the Schwarzschild radius — inside which the escape velocity exceeds the speed of light. For 110 years, every piece of evidence we have had about black hole horizons has been indirect. We have measured stars orbiting them. We have imaged the shadow they cast on hot gas. We have heard them collide. What we have never done is observe the horizon itself.

On 14 January 2025, the LIGO interferometers in Hanford, Washington and Livingston, Louisiana recorded the gravitational-wave signal GW250114: two black holes of roughly 30 and 25 solar masses spiralling together and merging into a single, spinning Kerr remnant. The signal was extraordinarily loud — clean enough that what would have been noise in earlier detections was, here, signal.

For eighteen months it sat in the data. On 24 June 2026, Nature published the analysis. The team, working between Canberra, Waterloo, Pasadena and Madrid, reported finding inside the ringdown a feature predicted by Sizheng Ma and colleagues' 2024 theoretical work but never observed: a "direct wave" — gravitational radiation oscillating at frequency 2Ω_H, where Ω_H is the rotation frequency of the newly formed black hole's horizon, and damping at a rate set by the horizon's surface gravity κ.

In plain language: the team has heard the horizon spin and felt how hard it pulls.

What the measurement actually is

A black hole's event horizon, in classical general relativity, is characterised by two numbers. Its angular velocity Ω_H tells you how fast the horizon rotates — which, because of an effect called frame dragging, sets the orbital rate of anything that falls in. Its surface gravity κ sets how rapidly signals from anything near the horizon get red-shifted into silence. Together, these are the horizon's complete classical fingerprint.

The Sun et al. paper reports a matched-filter signal-to-noise ratio of roughly 8 in the Hanford detector and 6 in Livingston for the direct-wave component — strong enough that the false-alarm probability is below one in a million. The measured values of Ω_H and κ agree, within their 90% credible intervals, with the predictions for a Kerr black hole derived from the inferred mass and spin of the merger remnant.

That second sentence is, depending on your temperament, either the most boring or the most extraordinary sentence in physics this year. General relativity, formulated by a man with no computer and no detector, predicts the precise oscillation frequency of a hole in spacetime that was created a billion light years away by an event that lasted a tenth of a second. We measured it. It was right.

The framework: what is being tested

The "no-hair theorem" is the formal name for the claim that an astrophysical black hole is completely specified by its mass and spin (and, in principle, charge — but realistic astrophysical black holes are uncharged). Every other property has to follow.

If GW250114 had shown an event horizon spinning at a slightly different rate than the no-hair theorem predicts — even by a few percent — we would be having a different conversation today. We would be entertaining "exotic compact objects": gravastars, fuzzballs, boson stars, firewalls, and the various stringy alternatives that have populated theoretical physics papers since the 1990s. Most of those alternatives predict horizon-replacement structures that should show up as deviations exactly here, in exactly this measurement.

They did not show up. The remaining room for new physics around the horizon is now measured, with this detection, at roughly 5% in Ω_H and 7% in κ. The next event of comparable loudness — and one is statistically expected within the current O5 observing run — will roughly halve those bounds.

This is how general relativity gets killed, if it gets killed: not by a thunderclap, but by a slowly tightening noose of measurements that refuse to misbehave. So far the noose is around general relativity's neck and general relativity is calmly drinking tea.

Who actually did this

The work is a useful illustration of where modern gravitational-wave physics actually happens. The lead authors are at the ARC Centre of Excellence for Gravitational Wave Discovery (OzGrav) at the Australian National University in Canberra. The theoretical framework was developed at the Perimeter Institute for Theoretical Physics in Waterloo, Ontario, with Yanbei Chen's group at Caltech, and Spanish collaborators at the IFAE in Barcelona and the IGFAE in Santiago de Compostela. The detectors are operated by the LIGO Scientific Collaboration, with data jointly analysed with the European Virgo collaboration and Japan's KAGRA.

There is no nationally heroic story here. Australia did not "discover" the horizon; nor did the United States or Spain or Canada. A genuinely globally distributed collaboration found a feature in jointly-owned data using publicly-developed software, and published it open-access. This is what late-stage scientific infrastructure looks like when it works.

Cross-layer implications

For fundamental physics: the strongest experimental constraint to date on alternative-gravity theories. Modified gravity models that predict horizon-replacement structures (Mathur-style fuzzballs, Mazur-Mottola gravastars, certain string-theoretic black hole mimickers) are now tightly constrained, and several specific parameter regions are excluded.

For black-hole thermodynamics: the measurement of κ is, by the laws of black-hole mechanics, also a direct measurement of the remnant's Hawking temperature — although at a thermodynamic temperature of order 10⁻⁸ kelvin for a stellar-mass black hole, the Hawking radiation itself remains far below any detection threshold. Still, the temperature is no longer a theoretical inference. It is a measured quantity.

For the next generation of detectors: the Cosmic Explorer (US) and Einstein Telescope (Europe) projects, currently in design phase, will detect tens of thousands of events at SNR > 100 per year. The horizon-direct-wave technique, which required a once-a-decade-loudness event with the current detectors, will become routine. Funding decisions on both facilities are expected in 2027–2028; this paper is now part of the case for both.

For the public imagination: the result will be reported as "scientists hear a black hole's surface" — and that framing is, for once, almost accurate. The surface of no return has just produced its first observational autograph.

The quieter story

The Sun-Lu paper relies on a theoretical framework developed by Sizheng Ma, then a postdoc at the Perimeter Institute, in a 2024 Physical Review D paper that was largely ignored at the time. Ma's predicted "direct wave" was treated as an interesting toy calculation. Eighteen months later it became the analytical key to extracting the strongest test of general relativity ever performed.

This is how theoretical physics tends to actually work: a single quiet paper, well-aimed, sitting on the arXiv for two years before the right data arrives. It is not the dominant narrative of how science makes progress — the dominant narrative is industrial-scale collaboration — but it is still how a lot of the most consequential individual contributions land.

Uncertainty ledger

• Single event. This is one detection. Replication on a second comparable-loudness event is the next milestone.
• Waveform modelling assumptions. The direct-wave extraction depends on a specific theoretical model. Alternative models (e.g. effective-one-body variants with different horizon assumptions) will reanalyse the data over the coming months and may revise the credible intervals.
• Noise systematics. The Livingston detector had a known scattered-light glitch episode in the hour before GW250114; the team accounts for it, but it will be re-checked.
• Not a test of quantum gravity. This measurement constrains classical horizon properties. The information paradox, firewall debates, and other quantum-gravitational questions about horizons are not addressed by this work.

What this means for the natural audience

For physicists and graduate students working on gravitational waves: the GW250114 strain data is publicly available through the Gravitational Wave Open Science Center. The Sun-Lu analysis code is published with the Nature paper. Reproducing the direct-wave extraction is now a realistic graduate-level project.

For astronomy communicators and educators: this is the first horizon measurement. That phrasing is correct and worth using carefully. It is not the first observation of a black hole, nor the first test of general relativity in the strong-field regime — both of which have substantial earlier literature.

For the general science-engaged public: the headline that will run today and tomorrow is some variation of "black hole's edge measured directly for the first time". That is accurate. The deeper sentence is: an Austrian physicist's notebook calculation from 1916 just got measured, again, and was again correct, this time inside the horizon.

For practitioners in adjacent fields: no immediate engineering or applied consequences. There will be no horizon-based technology in your lifetime.

Bottom Line

A 110-year theoretical edifice — Einstein's general relativity, Schwarzschild's horizon, Kerr's rotation, the no-hair theorem — has just been measured at the one place it was always supposed to break, and it didn't break. We have heard the surface of a black hole ring. That is the actual sentence. Anyone telling you otherwise is selling you a softer one because they don't trust you with this one.

Sources

• Nature — Sun, L., Lu, N., Ma, S. et al. "GW250114 reveals signatures of post-merger black-hole horizon" — 24 Jun 2026 (Tier 1, peer-reviewed)
• ScienceAlert — "Scientists May Have Detected The First Signature of a Black Hole's Event Horizon" — 24 Jun 2026 (Tier 2)
• Space.com — "Black hole's 'point of no escape' studied with the loudest gravitational waves ever heard" — 26 Jun 2026 (Tier 2)
• LIGO Scientific Collaboration / OzGrav official statement — 24 Jun 2026 (Tier 1, primary)
• Space Daily — analysis summary — 30 Jun 2026 (Tier 3, contextual)