The Bullet Cluster is probably the most famous single image in dark-matter science. Two galaxy clusters slammed into each other roughly 3.8 billion light-years away, and the collision ripped the clusters' hot gas away from their dark matter, leaving the mass (mapped through gravitational lensing) and the ordinary matter (mapped through X-rays) sitting in visibly different places. It's the closest thing astrophysics has to a smoking gun for dark matter as a real substance rather than a tweak to gravity.
But for 20 years, a more basic question about the same object has refused to settle down: how big were the two colliding clusters relative to each other? Was this a fender-bender between a heavyweight and a much smaller intruder, or something closer to a collision between equals? Published answers have ranged from roughly 2-to-1 up to 10-to-1 or beyond, depending on the telescope, the data, and the modeling choices behind each study.
A new analysis combining James Webb Space Telescope imaging with wide-field data from the Dark Energy Camera (DECam) says it has now nailed the number: 10.14, with an uncertainty of +3.22/-2.47. The main cluster weighs in at roughly 15.11×10^14 solar masses; the smaller subcluster, about 1.49×10^14 solar masses. In plain terms, that's about 1.5 quadrillion solar masses squaring off against about 150 trillion — a mismatch of an order of magnitude, not a fair fight.
The work, led by Boseong Young Cho of Yonsei University along with M. James Jee, Hyungjin Joo, Sangjun Cha, and Kim HyeongHan, was submitted to arXiv in December 2025 and has now been published in The Astrophysical Journal. It was also spotlighted by the American Astronomical Society's AAS Nova on July 6, 2026, and covered by Tech Times the following day.
Why did this take 20 years to pin down?
Gravitational lensing — measuring how background galaxies' light is distorted by the mass in front of them — is the standard tool for weighing galaxy clusters, since most of that mass is dark and invisible to every other kind of observation. But lensing comes in two flavors, and each has a weakness that matters here.
Strong lensing produces dramatic, unmistakable arcs and multiple images near a cluster's core, and it's precise — but only in that small central region. Weak lensing uses the subtle, statistical stretching of thousands of background galaxies across a much wider area, which makes it useful for reaching a cluster's full extent, but the signal is noisy and the resulting mass models can be biased depending on how far out the data extends and how the model is constructed.
The Bullet Cluster's mass has mostly been measured before with one method or the other, on relatively narrow fields of view. That combination — narrow imaging plus a single lensing regime — tends to force astronomers to extrapolate mass estimates beyond the region they actually observed, which is exactly where estimates have historically diverged.
What did the new study do differently?
Three things, according to the paper and AAS Nova's summary of it.
First, the team used JWST's NIRCam instrument to perform both weak and strong lensing analysis on the same data set, letting the precise strong-lensing measurements in the core correct for known biases in the weak-lensing model — rather than treating the two methods as separate, independent estimates to be reconciled after the fact.
Second, they paired that Webb data with wide-field imaging from DECam, a large-format optical camera on the Blanco 4-meter telescope in Chile. DECam's wide field of view meant the researchers could map weak lensing signal all the way out to the cluster's virial radius — the outer boundary within which a cluster's gravity holds itself together — instead of extrapolating a mass profile derived from a smaller, JWST-only footprint.
Third, and perhaps most importantly for solving the two-decade puzzle, the team identified three separate mass halos in the system rather than the two that simpler models had assumed: two distinct concentrations within the "main" cluster, plus one in the smaller subcluster. Earlier studies that fit the system as a clean two-halo collision were, in effect, missing structure that skewed their mass ratios.
Why It Matters
The Bullet Cluster isn't just a pretty picture — it's a benchmark. Its 2007 computer simulation by Volker Springel and Glennys Farrar became one of the most cited efforts to reconstruct exactly how the collision unfolded, including the impact velocities and geometry needed to explain the observed separation between gas and dark matter. That simulation assumed a roughly 10-to-1 mass ratio because the collision dynamics required it — a minor merger, where a smaller cluster plows through a much larger one at high speed, behaves very differently from a near-equal, head-on collision between two comparably sized clusters. A 2-to-1 encounter would produce different shock structures, different velocities, and a different story about how much time has passed since impact.
Until now, that 10-to-1 assumption was essentially an educated guess baked into the model to make the physics work, still awaiting direct confirmation from the mass measurements themselves. The new JWST-DECam result — 10.14, with error bars that comfortably exclude anything near 2-to-1 — closes that loop. It means one of the foundational simulations underpinning the modern understanding of the Bullet Cluster wasn't just internally consistent, it was right, and the measured ratio lines up with what LambdaCDM cosmology predicts for high-velocity mergers of this kind.
More broadly, the study is a demonstration of method as much as result. Combining Webb's resolving power with a wide-field ground-based survey camera, and using strong lensing to keep weak lensing honest, is a template other groups can apply to the dozens of other merging clusters whose mass ratios are similarly contested. Two decades of disagreement over one of astronomy's signature objects didn't get resolved by a single better telescope — it took stitching two different instruments and two different lensing techniques into one consistent picture.
