Scientists Have Dismissed Black Holes as Dark Matter for Years, New Math Says They Were Wrong

About 85% of the matter in the universe is invisible, and for years physicists have leaned on one big excuse to dismiss primordial black holes as an explanation: the idea seemed too “fine-tuned,” too dependent on everything lining up perfectly. A new study says that excuse doesn’t hold up once the math is actually checked.

Physicist Stefano Profumo of the University of California, Santa Cruz built a mathematical test for exactly how delicately balanced a dark matter theory needs to be to work, then ran it on 16 different theories, including black hole formation scenarios and popular particle candidates like WIMPs, using the same rules for every single one. The result, published in Physical Review D, flips the usual assumption: some black hole models are just as mathematically stable as the most trusted particle theories, while certain WIMP models are far shakier than they get credit for.

That result matters. It undercuts a long-standing argument that has shaped how physicists weigh primordial black holes against particle candidates. If black holes aren’t uniquely “too fine-tuned” after all, the search for dark matter may need a fresh look.

How Do Scientists Measure Dark Matter Fine-Tuning?

To understand this study, it helps to know what scientists mean by “fine-tuned.” A theory is fine-tuned if it only works when its numbers land in a very narrow, specific range, similar to a combination lock that opens at one exact number out of a million. If a tiny change to that number breaks the theory entirely, scientists get suspicious. A theory that still works across a wide range of values is considered “natural.”

Profumo used a decades-old formula called the Barbieri-Giudice measure, originally built to test whether certain particle physics ideas were believable. It asks a simple question: if a model’s input number shifts slightly, how much does the predicted amount of dark matter change? A small nudge causing a small change means the theory is well-behaved. A small nudge causing a massive swing means the theory barely holds together. Importantly, this captures one specific, mathematical kind of fine-tuning. A high score flags a theory whose prediction is touchy to its inputs, but it is not a final verdict on how believable that theory is overall, a fuller judgment the paper deliberately leaves open.

Profumo applied this test across 16 different dark matter models, then ranked 12 representative benchmark setups side by side. That lineup covered three ways black holes could form without cosmic inflation (the theorized rapid expansion of the early universe), six inflation-based black hole scenarios, and seven particle candidates, among them different versions of WIMPs, a feebly interacting particle called a FIMP, asymmetric dark matter, and two versions of a hypothetical particle called the axion. Every scenario was measured against the same target: the actual amount of dark matter astronomers have observed in the universe, pinned down by the Planck satellite mission.

This wasn’t a test involving new telescope data or lab experiments. It was a mathematical audit. Profumo worked through the equations behind each theory, calculated how sensitive each one is to small changes in its inputs, and compared the results using the same rules throughout. He also compared the results against two related sensitivity measures, which generally tracked the same pattern.

What the Dark Matter Numbers Revealed

Profumo’s rankings split the twelve scenarios into three tiers, based entirely on the mathematical shape of each theory’s formula, not on whether the dark matter in question was a particle or a black hole.

Sitting in the most “natural” tier were asymmetric dark matter, one version of the axion, a type of black hole formed from structures called biased domain walls, an “off-resonance” WIMP, and the FIMP particle. These theories shared a key trait: their predicted dark matter amount scaled straightforwardly with their inputs, without any hidden explosive sensitivity.

A middle tier included black holes formed during a hypothetical early era when matter dominated radiation, along with a WIMP model that depends on two similarly sized particles interacting to shape the final result. Both landed in the middle for the same reason: their formulas both come down to a single exponential term tied to how the universe cooled over time.

At the opposite end sat the most fine-tuned tier, and here the story turned dramatic. One WIMP model relies on hitting a single exact “resonance” condition, much like tuning a radio to one precise frequency. That makes it extremely delicate. Black holes born from a sudden, violent event called a first-order phase transition proved just as touchy, once Profumo swapped in a more accurate formula for how they form. Worst of all was the black hole recipe scientists discuss most often, tied to a phase of cosmic inflation called ultraslow-roll, when the field driving inflation briefly slows to a crawl. That scenario came out as the most fine-tuned in the group, in some cases by many orders of magnitude worse than the friendliest black hole models.

Profumo notes in the paper that the widely repeated claim about primordial black holes lumps together the worst-case scenario with a full range of possibilities spanning every level of naturalness.

The Bottom Line on Dark Matter Theories

This study doesn’t prove dark matter is made of black holes, and it doesn’t rule out WIMPs, axions, or any other candidate. What it does is remove a lazy argument from the table. Saying “black hole dark matter is too fine-tuned,” without specifying which formation method and comparing it fairly to equally fine-tuned corners of particle physics, doesn’t hold up once the math is actually done side by side. Some black hole scenarios are just as comfortable, numerically, as the most beloved particle models. Some particle models carry hidden tuning problems of their own. The fairest way to judge any dark matter theory, it turns out, is the way this paper did it: line up the equations, apply the same standard, and let the numbers do the talking instead of reputation.