The solar system’s giant planets break apart in 9 of 10 simulations

Jupiter, Saturn, Uranus and Neptune have long been treated as the most durable machinery in the solar system, a set of orbits expected to outlast the universe many times over. New simulations say the Sun itself will break that machinery as it dies, most likely within about a billion years of turning into a white dwarf.

The culprit is the way the Sun loses weight. Earlier calculations assumed a dying Sun would shed its outer layers smoothly, like a slow leak, which stretches every orbit without disturbing the order of the planets. The new work, by Konstantin Batygin and Jim Fuller of Caltech and Fred Adams of the University of Michigan, starts from evidence that dying stars throw off their envelopes in separate, lopsided bursts. Each burst gives the Sun a small shove, and the planets feel every one of them.

Across hundreds of simulated futures, those shoves added up. In the scenario the authors consider closest to the observations, about 40% of the simulated solar systems were already disrupted or violently scrambled by the time the white dwarf formed, and roughly nine in ten fell apart within 3 billion years. In some runs Saturn was thrown out of the solar system altogether. In others, Uranus and Neptune swapped places.

If the result holds, the end of the solar system stops being a remote accident delivered by a passing star. It becomes part of how the Sun dies, and the same logic reaches almost every planetary system in the galaxy, because nearly all stars end their lives as white dwarfs.

Why a dying Sun kicks its own planets

In roughly 5 billion years the Sun will swell into a red giant, and over its giant phases it will lose about 46% of its mass. With less mass holding them, the planets drift outward. In the smooth picture every orbit grows by a factor of about 1.85, so the outer solar system becomes a larger, slower copy of the one we know.

The hint that the loss is not smooth comes from Gaia, the European Space Agency’s star-mapping satellite. By tracking pairs of stars that orbit each other at great distances, astronomers found that newborn white dwarfs move as if they had been kicked at about 0.75 km per second, roughly 2,700 km/h. A star that shed its gas evenly in every direction would get no kick at all.

The leading explanation, a framework Fuller recently proposed, is that a red giant’s envelope leaves in discrete plumes driven by the churning of its outer layers. Telescopes that combine light from several mirrors already show separate, shifting dust clouds around aging giant stars. Every plume pushes the star slightly in a random direction, and the pushes add up to the observed kick.

For the planets, a shove to the Sun is a shove to every orbit at once. In the study’s reference case, the Sun receives about 4,600 recoils of a few metres per second each, with a median of about 7 m/s. Individually they are tiny. Because they strike at random points in each planet’s orbit, they do not cancel out, and the orbits start to wander. “Everything changes when the envelope departs in pieces,” the authors write.

How they did it

The team used rebound, a widely used open-source code that computes how bodies move under each other’s gravity, and paired it with a model of the Sun’s evolution built with MESA, a standard stellar-evolution program. The four giant planets started from their present positions and velocities, and the Sun then lost mass along that model track.

First came the control experiment. In more than 60 runs with smooth mass loss, the outcome never changed: the orbits expanded, the ratios between the planets’ orbital periods stayed frozen, and the system remained stable for billions of years.

Then the researchers broke the mass loss into pieces. They ran 672 simulations, varying the size of each ejected parcel of gas from one ten-millionth to one ten-thousandth of the Sun’s mass, and varying whether the bursts ran through all of the Sun’s giant phases or only the last one, called the asymptotic giant branch. Every end state was then followed for another 3 billion years with no passing stars or other outside disturbances.

From 10¹⁸ years to about one billion

The numbers being replaced are enormous. Chaos alone was estimated to need about 10¹⁸ years, a billion billion years, to destabilize Uranus. Adding passing stars, which find the widened orbits an easier target, cut the estimate to between 30 and 100 billion years. The universe is about 13.8 billion years old.

At the calibration the authors favour, the solar system now unravels on its own about a billion years after the white dwarf forms, a hundred times sooner than the expected wait for the first disruptive passing star. In the most violent runs, Neptune’s orbit was stretched to an eccentricity of about 0.5, so its distance from the Sun would swing by a factor of three between its closest and farthest points.

The trouble does not always wait for the Sun to finish dying. In 37 of 48 runs of the favoured scenario, orbits began to cross before the mass loss was over, the earliest while the Sun was still a red giant holding 89% of its present mass. In one example, Saturn was ejected within 2.5 million years of the white dwarf’s birth, leaving Uranus and Neptune locked together.

Isaac Newton suspected that the mutual pull of Jupiter and Saturn would eventually wreck the planets’ order. Laplace and Lagrange later showed that, to their level of approximation, the orbits only oscillate, and three centuries of calculations kept pushing the verdict further away. “Newton’s envisioned instability is real after all. He was mistaken only about the perpetrator,” the authors conclude.

Not every outcome was destructive. Some runs trapped pairs of planets in resonances, rhythmic orbital patterns the solar system has never had: Jupiter and Saturn in a 5:2 ratio, or Uranus completing two orbits for every one of Neptune. In 8 of the 672 simulations both pairs were trapped at once. Those locked systems were no safer, and in the most granular runs they survived slightly less often than the others.

What the simulations do not settle

Everything hinges on a quantity nobody has measured directly: how large each puff of escaping gas is. If the Sun sheds its envelope in parcels of a ten-millionth of its mass or less, the bursts largely average out and the solar system survives as an enlarged copy of itself. The violent outcomes take over once parcels reach about a hundred-thousandth of a solar mass.

The bridge from the Gaia kicks to parcel size is also an interpretation. It assumes the kicks come from many independent ejections. The authors argue that even the low end of the measured range, 0.25 km per second, implies parcels already in the violent regime. They also flag that their main scenario extends the bursts beyond the asymptotic giant branch, the phase where the model is best calibrated. Limiting the bursts to that phase removed every early break-up during the red giant stage, yet 83% of the systems still went unstable within 3 billion years.

The work leaves the inner planets aside. It follows only the four giants, so Mercury, Venus, Earth and Mars are not part of the calculation. Earlier studies put Mercury’s own risk at about 1%: that is the chance its orbit becomes unstable before the Sun leaves its current, hydrogen-burning phase.

The theory also makes predictions astronomers can check. If dying stars routinely shake their planets loose, young white dwarfs should travel with a population of free-floating planets. In these simulations the giants were far more often ejected than smashed together, and the debris that does fall onto white dwarfs should come mostly from rocky bodies rather than giant planets, which matches the rocky material measured in polluted white dwarfs so far.

Common questions about the Sun’s death and the planets

Will the solar system fall apart when the Sun dies?

According to these simulations, probably yes for the giant planets. If the Sun loses its outer layers in thousands of separate bursts, as white dwarf kicks suggest, roughly nine in ten simulated systems became unstable within 3 billion years of the white dwarf forming. With smooth mass loss, the planets stayed stable.

When will the Sun become a red giant?

In roughly 5 billion years. The Sun will then lose about 46% of its mass over its giant phases before ending as a white dwarf of about 0.54 solar masses. In the study, the break-ups happen during that mass loss or in the following few billion years.

What is a white dwarf kick?

It is the small velocity a star appears to gain as it becomes a white dwarf, about 0.75 km per second according to Gaia measurements of wide pairs of stars. A kick means the star lost its gas unevenly, and each uneven burst also nudges any planets in orbit.

Could Saturn really be thrown out of the solar system?

In some simulations, yes. In one run Saturn was ejected within 2.5 million years of the white dwarf forming. Among the giant planets, ejections were more common than collisions, so lost planets would drift through the galaxy as free-floating worlds.

What comes next

The study appeared in The Astrophysical Journal Letters in September 2026, after a preprint on arXiv, and drew wider coverage on 12 October. Its central number can be sharpened with more Gaia measurements of white dwarf kicks and with better models of how giant stars eject gas. The authors also point to the planets and debris now being found around white dwarfs as a record of how each star died, so every new system counted there tests whether the Sun’s giants are doomed or merely enlarged.

Reference: Batygin et al., “Terminal Instability of the Solar System Triggered by Stochastic Solar Mass Loss,” The Astrophysical Journal Letters, 2026. DOI: 10.3847/2041-8213/aea290

Tags: astronomy, white dwarf, sun, Caltech, Konstantin Batygin, solar system