Gravitational waves expand their horizons

Gravitational waves (GWs) have been part of our observational compendium for ten years and counting, and with the recent release of the fifth Gravitational-Wave Transient Catalog (GWTC-5), they appear to be entering a new phase of maturity. Published on 26 May 2026, GWTC-5 adds 172 binary merger events to the previously known 218 and so almost doubles LIGO–Virgo–KAGRA’s (LVK’s) contribution to the stellar graveyard1,2. It comes less than a year after GWTC-4 doubled the previous count, and will be superseded again in December by the release of data from the third and final part of the same observing run. But whereas the first GW detection prompted televised press events and celebrations across the world, GWTC-5 came much more quietly, perhaps reflecting that binary black hole (BH) mergers have now been thoroughly incorporated into the astrophysical canon. One might compare this trajectory to that of exoplanetary science, which began with a trickle of exoplanet discoveries in the 1990s but only established itself as a data-driven field — and as an independent field in its own right — over the next two decades, when the Kepler Space Telescope caused detections to flood in. The three back-to-back GWTC releases may mark a similar transition for GW science, though the fast approach of new technology means this new status quo will almost certainly be short-lived.

Notable in the evolution of exoplanet studies is that the novelty value of individual detections was quickly equalled by the scientific value that larger datasets allowed: modelling demographics, characterizing atmospheres, understanding planetary formation and more. With GWTC-4 and now GWTC-5, we are also seeing a shift in focus, with the few-hundred-strong samples injecting momentum into a range of subfields, among which black hole population studies stands out as having made particular progress. Certain individual detections still offer a high scientific yield — such as the loudest-yet event, GW250114, enabling tests of BH physics3 — but the strength in numbers means this is no longer the only mode of study available.

Though population analyses of GWTC-4 are still fresh, they seem to be converging on evidence for two (or more) BH subpopulations4,5,6,7, each with distinct mass and spin distributions and therefore likely coming from independent formation channels. An article in this issue presents one such analysis, in which Fabio Antonini and colleagues identify the higher-spin, broader-mass-range subpopulation as potentially having formed through the hierarchical merging of an earlier generation of BHs. Interestingly, a steep decline in the merger rate at around ~40 M ☉ separates the two subpopulations and may signal the presence of a mass gap in the low-spin, putatively first-generation subpopulation. In theory, such a mass gap might be carved out by pair-instability supernovae (PISN), which leave no remnant, but other models of the GWTC-4 data dispute whether this gap is really present8. The addition of GWTC-5 reasserts the existence of a second high-spin subpopulation but seems to weaken evidence for the mass gap9. Undoubtedly multiple independent studies will help clarify this picture, but for the time being, it looks like the doubling of the dataset may not redefine our understanding of the processes at play here, and that gradual developments in analysis techniques will also be important.

The December data release will be followed by a pause on the data front, as all four GW detectors have been undergoing upgrades since the end of 2025 and are not due back in operations until late this year. At that point, we can look forward to an intermediate six-month observing run to coincide with the start of LSST operations, prior to the beginning of the fifth main observing run some time in 2029. The upgrades alone are expected to push the field into new throes of discovery — helped along by the addition of LIGO-India, whose distance from the existing detectors will substantially improve source localization, around 2031 — but by then the GW community will also be preparing for the first generation of space-based detectors. The Laser Interferometer Space Antenna (LISA), TianQin (天琴计划) and B-DECIGO are all planned for launch within the next decade, and their similar experimental setups, each comprised of three orbiting spacecraft linked interferometrially over enormous distances, will enable the detection of a plethora of as-yet-inaccessible GW sources. Furthermore, the sensitivities of all three will mean that thousands to millions of disparate events will be present in any individual dataset (and will therefore need to be modelled simultaneously). The next generation of ground-based detectors — the Einstein Telescope and Cosmic Explorer — have more nebulous timescales, but should also become operational in the late 2030s. Surely this throwing-open of the parameter space will sweep GW science off whatever feet it is currently finding.