Into the Fog: What Dark Matter Could Be and How We Will Know

The WIMP era is ending in a haze of solar neutrinos. Here is what dark matter could still be, and the experiments that could finally tell us.

In December 2025, the LZ experiment, a seven-tonne tank of liquid xenon buried a mile underground in a former South Dakota gold mine, announced the most sensitive dark matter search ever performed: 417 live days of data and not a single confirmed dark matter particle (SLAC, 2025). The same analysis contained a genuine first. LZ detected boron-8 neutrinos streaming from the sun’s core, recoiling off xenon nuclei in exactly the way a light dark matter particle would. The detector had become so quiet, and so large, that the sun itself is now the background.

Physicists call this the neutrino fog (O’Hare, 2021). It is the point at which neutrinos, which no amount of rock can shield against, begin to mimic the signal that three generations of detectors were built to find. The December result was a milestone and an obituary at once: the flagship strategy of the past forty years is approaching its floor without a discovery.

Six months later, on the last day of June 2026, the Vera C. Rubin Observatory in Chile opened the other front. Its ten-year Legacy Survey of Space and Time will photograph the entire southern sky every few nights, mapping the gravitational fingerprint of dark matter across billions of galaxies (Rubin Observatory, 2026). Between these two events sits the current state of the field: never more certain that dark matter exists, never less certain what it is (Garisto, 2026).

This essay is a map of that uncertainty. What the candidates are, which ones are gaining ground, and, more importantly, what evidence could actually settle the question in the next two decades.

The evidence that will not go away

It is worth restating why nobody serious doubts the phenomenon, even as every candidate explanation struggles. The evidence is gravitational, independent, and spans a factor of a billion in physical scale.

Zwicky noticed in 1933 that galaxies in the Coma cluster move far too fast to be held together by the mass we can see (Zwicky, 1933). Rubin and Ford found in 1970 that stars at the edges of galaxies orbit as if embedded in a vast halo of unseen mass (Rubin & Ford, 1970). The Bullet Cluster showed the unseen mass and the visible gas literally separated from each other during a collision, which is very hard to fake with modified gravity alone (Clowe et al., 2006). And the cosmic microwave background encodes the ratio precisely: about five sixths of the matter in the universe is dark (Planck Collaboration, 2020).

All of this tells us how much there is and how it clusters. It tells us almost nothing about what a single piece of it looks like. The allowed mass range for the constituent, whatever it is, spans roughly ninety orders of magnitude, from particles lighter than any known quantum of energy to black holes heavier than asteroids (Bertone & Hooper, 2018). The story of the last decade is the slow, expensive elimination of the middle of that range, and the migration of theorists toward its exotic edges.

The end of the WIMP century

The weakly interacting massive particle was never just a guess. It fell out of supersymmetry, the most elegant proposed extension of the standard model, and it came with a numerical miracle: a particle with roughly the mass of a heavy atom and the weak nuclear force’s interaction strength would be produced in the early universe in almost exactly the abundance dark matter requires. Two problems, one particle.

That coincidence drove the LHC and the xenon program. Both came up empty. The LHC found no superpartners, and LZ’s December 2025 limits, the strongest ever set, now press directly into the neutrino fog at low masses (SLAC, 2025). The community’s answer was to be one final experiment: XLZD, a 60 to 80 tonne xenon observatory that would search essentially all the remaining WIMP parameter space down to the fog itself (XLZD Collaboration, 2025). Its future is now genuinely uncertain. At the end of 2025 the US Department of Energy declined to host or fund its share of the project, and members of the collaboration concede it may not happen at all (Garisto, 2026).

Whatever XLZD’s fate, the intellectual shift has already happened. The WIMP is no longer the default, and the field has stopped behaving as if dark matter owes us a particle that is convenient to detect.

The axion’s decade

If the WIMP was the candidate of the 1990s and 2000s, the axion owns the 2020s. Like the WIMP it was not invented for dark matter. It emerged in 1977 as a by-product of the Peccei-Quinn mechanism, which explains why the strong nuclear force treats matter and antimatter identically when nothing in the theory says it must (Peccei & Quinn, 1977). An axion light enough to be dark matter would behave less like a particle and more like a classical field oscillating everywhere in space, with a mass somewhere between a trillionth and a millionth of an electron’s.

Detection is a radio engineering problem. Inside a strong magnetic field, an axion can convert into a photon whose frequency reveals the axion’s mass. Haloscope experiments tune ultracold resonant cavities across frequency bands like scanning a dial, listening for a tone that never switches off. In late 2025 ADMX reported its deepest scan yet, covering 1.1 to 1.3 GHz at sensitivity to the benchmark KSVZ axion model (ADMX Collaboration, 2025). HAYSTAC, ORGAN, MADMAX and a growing family of experiments with increasingly theatrical acronyms are attacking neighbouring bands, while DMRadio targets the lighter, GUT-scale axions below a microelectronvolt.

The honest accounting is that only 10 to 20 percent of the parameter space for a strong-CP-solving axion has been searched (Garisto, 2026). But unlike the WIMP, the axion program has a defined finish line. The technology improves each year, quantum amplifiers now operate near the fundamental noise limit, and the plausible mass window can realistically be scanned end to end within the 2030s. The axion is that rare thing in this field: a hypothesis with a scheduled falsification date.

Lighter and stranger: the hidden sector

Beyond the two classic candidates, theory has diversified into what is loosely called the dark sector: the idea that dark matter is not one particle bolted onto our standard model, but the visible tip of its own private physics, with its own forces and possibly its own spectrum of particles. Mirror-world models, atomic dark matter, and recent two-component proposals in which detectable signals arise only from interactions between dark matter species all live here (Garisto, 2026).

One version of this idea has real astrophysical motivation. Self-interacting dark matter, in which dark matter particles scatter off each other but not off us, was proposed to explain why the cores of some dwarf galaxies look less dense than cold dark matter simulations predict, while some cluster substructures look denser (Tulin & Yu, 2018). A single extra interaction can produce both behaviours, since self-interactions first flatten a halo’s core and later trigger its gravothermal collapse. New simulation techniques developed in the past year make these models properly testable against observation for the first time.

The experimental counterpart is the low-mass frontier: candidates between an electron and a proton in mass, too light to kick a xenon nucleus hard enough to notice. Detecting them means registering the energy of a single ionised electron or a quantised vibration in a crystal lattice. The new generation of detectors looks nothing like the giant xenon tanks. They are tabletop devices: skipper CCDs that count individual electrons, sapphire and gallium arsenide crystals read out by superconducting sensors, and a planned experiment at Modane in France built around roughly a tablespoon of superfluid helium, in which a dark matter impact would spray helium atoms onto silicon sensors above (Garisto, 2026). The struggle at this scale is noise. Detectors sensitive to single quanta are sensitive to everything, and the community is still learning what its own backgrounds are, sometimes the hard way: one apparent signal turned out to be a crystal clamped too tightly.

Wave dark matter and the smallest scales

At the extreme light end sits fuzzy dark matter, an ultralight boson around 10^-22 eV whose quantum wavelength is measured in kiloparsecs. A galaxy’s worth of it behaves like a single coherent wave, suppressing structure below a characteristic scale and forming solitonic cores at the centres of halos (Hui, 2021). This is an appealing idea because it is so visibly different: the astrophysics itself becomes the detector.

It is also an idea in trouble, at least in its purest form. The Lyman-alpha forest, the thicket of absorption lines that quasar light picks up passing through intergalactic hydrogen, measures the clumpiness of matter on exactly the scales fuzzy dark matter smooths away. The latest analyses disfavour the canonical mass range and push the allowed particle mass higher, or force fuzzy dark matter to be only a fraction of the total (Liu et al., 2026). Theorists have responded with self-interacting variants whose solitons are more compact and which may evade the small-scale constraints. The next few years of dwarf galaxy kinematics and lensing data will decide whether that is a rescue or an epicycle.

The heavyweight option

The one candidate requiring no new particle physics at all is the primordial black hole: black holes formed not from stars but from overdense patches in the first second after the Big Bang. Decades of constraint-building have closed most of the mass range, but a window survives around asteroid masses, roughly 10^17 to 10^22 grams, where such objects are too light to lens stars detectably and too heavy to have evaporated (Green & Kavanagh, 2021). Recent work keeps chipping at the edges, with new limits arriving from the thermal history of intergalactic gas, from pulsar timing, and from high-energy neutrino searches.

What makes this corner of the field feel alive is the inventiveness of the proposed probes. One 2025 paper points out that Ganymede’s icy shell is a recording medium: a primordial black hole punching through would leave a distinctive wound, refrozen from the subsurface ocean, unlike any asteroid crater (DeRocco, 2025). Others propose reading dark matter annihilation out of planetary atmospheres as a faint ultraviolet airglow (Blanco et al., 2024). The solar system, it turns out, is a detector array that has been integrating signal for four billion years.

The gravity-only endgame

There is a sharper way to frame the field’s predicament: every single thing we know about dark matter, we learned through gravity. All the null results constrain the optional extras, the couplings dark matter might have to our particles. A growing school of thought, associated with theorists like Kathryn Zurek, argues the search should be re-anchored to the one interaction that is guaranteed to exist (Garisto, 2026).

That programme is concrete, if daunting. The Windchime project proposes an array of billions of quantum-limited mechanical sensors that would feel the gravitational tug of a single Planck-mass dark matter particle passing through, a detection requiring no non-gravitational coupling whatsoever (Windchime Collaboration, 2022). MAGIS-100, a 100-metre atom interferometer being commissioned at Fermilab in 2027, drops clouds of strontium atoms in vacuum and reads phase shifts that ultralight dark matter would imprint across twelve orders of magnitude in mass (Abe et al., 2021). Pulsar timing arrays are beginning to constrain how dark matter clumps on sub-galactic scales, where we currently know almost nothing. Zurek’s own estimate for the pure-gravity programme is measured in decades, possibly a century. It is the patient, guaranteed-to-teach-us-something path.

Honesty requires mentioning the rival interpretation. Modified Newtonian dynamics still claims the low-acceleration regime, and the cleanest test, wide binary stars from Gaia whose internal accelerations fall below the MOND threshold, remains actively contested: one group reports a persistent 40 to 50 percent gravitational boost exactly where MOND predicts it (Chae, 2023), while independent analyses with more careful modelling of hidden triple companions find plain Newtonian gravity fits better (Pittordis et al., 2025). Gaia’s final data releases should make this test decisive. It is one of the few places in the entire subject where a clean yes-or-no answer is scheduled.

How we will actually know

Predictions in this field have a poor track record, so here is a checklist rather than a forecast. These are the concrete ways the question gets resolved, or transformed, in the next two decades.

A direct hit remains possible. If XLZD is built, it either finds WIMPs or retires the hypothesis into the neutrino fog (XLZD Collaboration, 2025). The haloscope network scans the remaining QCD axion window through the 2030s; an axion discovery would look like a persistent narrow tone in a cold cavity, and its absence across the full band would be nearly as informative. The tabletop experiments cover the electron-to-proton mass range in parallel.

Astronomy may get there first. Rubin’s decade-long survey will measure the halo mass function down to scales far below the smallest visible galaxy, using stellar streams, satellite counts, and strong lensing anomalies. Cold, warm, fuzzy, and self-interacting dark matter make cleanly different predictions there (Rubin Observatory, 2026). This is the rare probe that does not depend on dark matter interacting with us at all, only on how it clumps. Combined with Euclid’s complementary space-based lensing map, the small-scale structure data of the early 2030s may quietly rule out more candidate space than every underground laboratory combined.

And the long game is gravity: atom interferometers, mechanical sensor arrays, pulsar timing, and eventually detectors we have not designed yet, measuring the dark matter field through the only force it is guaranteed to feel.

There is a real possibility that none of these finds a particle, and that outcome would itself be profound. It would mean dark matter lives in a sector so hidden that gravity is its only handle, or that it is black holes from the first second of time, or that our theory of gravity fails in ways subtler than any current alternative. The fog LZ has sailed into is not the end of the map. It is the coastline of the part we had charted, and the fleet leaving it now is stranger, smaller, more numerous, and better instrumented than anything the field has launched before.

References