For more than a decade, the collision between the Milky Way and Andromeda galaxies has been treated as a settled fact of cosmic destiny. A peer-reviewed study published in Nature Astronomy now challenges that certainty. By adding the gravitational influence of the Large Magellanic Cloud and the Triangulum galaxy (M33) to orbital simulations, researchers found that the merger is no longer a foregone conclusion but a probabilistic outcome, with odds that fall well below the near-certainty that NASA announced in 2012.
Why the collision forecast has shifted from certainty to probability
The original case for an inevitable smashup rested on Hubble Space Telescope observations that pinned down Andromeda’s sideways motion across the sky for the first time. Those measurements, reported in a series of papers using multi-epoch HST data with background galaxies as reference points, showed that M31’s velocity vector was statistically consistent with a nearly radial approach toward the Milky Way. Orbit integrations and collisionless N-body simulations then projected a merger unfolding over roughly 10 billion years. NASA’s public statement at the time declared the Milky Way “destined for head-on collision” on a multi-billion-year timescale.
That projection, however, treated the Milky Way and Andromeda essentially as a two-body problem. The 2025 study reframes the question by including two additional massive objects in the Local Group: M33 and the Large Magellanic Cloud. Both exert time-varying gravitational tugs that alter the predicted trajectory of Andromeda relative to our galaxy. When those perturbations are folded into the simulations, the range of possible outcomes widens dramatically. Some runs still produce a direct collision. Others show the galaxies swinging past each other on a wide arc, never fully merging within the simulation window.
Hubble, Gaia, and the data trail behind the new model
The chain of evidence starts with the first absolute proper-motion measurement of Andromeda, which used years of Hubble imaging to detect the galaxy’s tiny lateral drift against distant reference objects. That breakthrough allowed astronomers to estimate Andromeda’s full three-dimensional velocity for the first time, rather than relying solely on the line-of-sight speed derived from redshift. Follow-up work modeled the future orbital evolution of the Milky Way, M31, and M33, producing the collision timeline and visualizations that entered popular science coverage worldwide.
Gaia DR2 data later refined those proper-motion estimates. Researchers used the European Space Agency’s astrometric satellite to re-measure the motions of both M31 and M33, tightening constraints on the transverse velocity and re-evaluating the orbital geometry. The updated velocity numbers shifted the predicted encounter from a near-certain head-on strike to something more glancing, though the exact outcome remained sensitive to assumptions about total mass and dark-matter halo shapes.
The recent Nature Astronomy analysis builds directly on that data trail. Its simulations sample a range of initial conditions drawn from the latest astrometric and kinematic constraints, then integrate the orbits forward while allowing M33 and the Large Magellanic Cloud to act as live gravitational perturbers rather than static background objects. The result is a distribution of outcomes rather than a single predicted trajectory, with some realizations ending in a long-delayed merger and others in a close but non-destructive flyby.
What the Large Magellanic Cloud changes in the calculation
The Large Magellanic Cloud is the Milky Way’s most massive satellite galaxy, and recent mass estimates have revised it upward significantly compared with older values. A heavier LMC shifts the Milky Way’s center of mass and alters the reflex motion of the galactic disk, which in turn changes the relative velocity between the Milky Way and Andromeda. Because the LMC is on its own evolving orbit, its gravitational pull is not constant but varies over the billions of years covered by the simulation.
M33 plays a similar role on Andromeda’s side. Its orbit around M31 introduces additional angular momentum into the system, and close passages between M33 and M31 can nudge Andromeda’s trajectory enough to change whether the two large spirals collide or merely interact at a distance. The combined effect of these two perturbers is large enough to push the merger probability below the near-100-percent confidence that earlier two-body models implied.
In the new work, the researchers treat all four galaxies as extended, self-gravitating halos rather than point masses. This allows the LMC and M33 to exchange angular momentum and energy with their hosts, subtly reshaping the Milky Way and Andromeda halos over time. Those long-term shifts feed back into the relative orbit of the two giants, making the system’s fate more sensitive to initial conditions than earlier, simpler models suggested.
Open questions and the next observational test
Several uncertainties remain. The total masses of the Milky Way and Andromeda are still debated, with different methods yielding ranges that affect the predicted closing speed. The dark-matter halo profiles of both galaxies are inferred rather than directly measured, and small changes in halo concentration or extent can swing outcomes between merger and flyby. The LMC’s own mass is constrained but not pinned down to the precision needed for a definitive answer.
One concrete way to test the 2025 results would be to re-run the N-body simulation suite with Gaia DR3 proper motions, which offer tighter astrometric precision than the DR2 catalog used in the underlying calculations. A recent methodological paper on Local Group dynamics outlines how improved astrometry can be propagated into orbit forecasts, providing a roadmap for updating the Milky Way–Andromeda scenario as new data arrive.
Additional constraints on the LMC and M33 masses will also be crucial. Stellar streams in the Milky Way halo, for example, act as sensitive seismographs of past gravitational perturbations. By modeling how those streams are bent and heated, astronomers can back out the strength of the LMC’s pull over time, narrowing the allowed mass range. Similar techniques applied to satellite systems around Andromeda may refine the mass of M33 and its orbital history.
Even with those improvements, the study argues that the Local Group’s future should be framed in probabilistic terms. Rather than a single, cinematic prediction of two disks colliding on cue, the new picture is a spectrum of possibilities: a prompt merger, a delayed coalescence after multiple passages, or a near miss that leaves both galaxies distorted but distinct. In each case, the night sky billions of years from now would look dramatically different from today, but the details will depend on small differences in mass and motion that current instruments are only beginning to resolve.
For now, the main conclusion is conceptual rather than catastrophic. The Milky Way is no longer described as unavoidably doomed to a specific collision; instead, it inhabits a dynamically rich environment where the fates of giant galaxies are shaped by their smaller companions. As new astrometric releases and deeper surveys refine the inputs, the odds on each outcome will be recalculated, turning what once seemed like destiny into a long-running experiment in cosmic statistics.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
