Worlds in Collision: How a Star's 'Completely Bonkers' Flickering Revealed a Cataclysmic Planetary Crash 11,000 Light-Years Away

Anastasios "Andy" Tzanidakis was sifting through old telescope data when he stumbled onto something that should not have been there. A sun-like star, 11,000 light-years away, had started behaving in ways that defied every expectation astrophysicists have for a stable, middle-aged sun. Its light was dimming, flickering, and surging in infrared — hallmarks, Tzanidakis and his colleagues now argue, of a cataclysmic collision between two entire worlds.

The discovery, published March 11, 2026 in The Astrophysical Journal Letters, represents one of only a handful of planetary collisions ever observed in real time [1][2]. But its significance extends far beyond the exotic: the debris cloud orbits at roughly the same distance from its star as Earth sits from ours, raising the tantalizing possibility that what astronomers are watching unfold is a replay of the violent event that created our own Moon.

A Star That Went 'Completely Bonkers'

The star in question, cataloged as Gaia20ehk, sits near the constellation Puppis. It is an F-type main sequence star — slightly hotter and brighter than our Sun, but fundamentally the same kind of stable, hydrogen-burning furnace that should produce steady, predictable light for billions of years [3][4].

When Tzanidakis, a doctoral candidate in astronomy at the University of Washington, first pulled up archived observations of Gaia20ehk in 2020, the light curve told a different story. "The star's light output was nice and flat, but starting in 2016 it had these three dips in brightness," Tzanidakis told reporters. "And then, right around 2021, it went completely bonkers. Stars like our sun don't do that" [1][2].

The dips were not subtle. Huge quantities of rocks and dust — seemingly materializing from nowhere — were passing in front of the distant star as the material orbited the system, patchily dimming the light that reached Earth. Something catastrophic had happened in the Gaia20ehk system, and multiple telescopes, spanning years of archival data, had inadvertently captured the aftermath.

The Smoking Gun: When Infrared and Visible Light Diverge

The critical piece of evidence came from comparing observations across different wavelengths. While the visible light from Gaia20ehk was flickering and fading, something extraordinary was happening in the infrared: the star was getting brighter [1][3].

"The infrared light curve was the complete opposite of the visible light," Tzanidakis explained. As the visible light began to flicker and dim, the infrared light spiked. This inverse relationship pointed to a single conclusion: the material blocking the star's visible light was not cold debris drifting through space. It was superheated — so hot that it was glowing intensely in the infrared [2][4].

A cataclysmic collision between two planets would produce exactly this signature. The unimaginable energies released in a planetary-scale impact — two worlds, each potentially comparable in mass to Earth, smashing together at velocities of tens of kilometers per second — would vaporize rock, melt silicates, and scatter a vast cloud of incandescent debris across the system. That cloud, still radiating heat, would glow brightly in infrared while simultaneously blocking visible starlight as it transited in front of the host star.

The research team, led by Tzanidakis with senior author James Davenport, a UW assistant research professor of astronomy, proposed a two-phase model for what they were observing [3][4]. The three initial brightness dips beginning in 2016 may represent grazing impacts — the two planets spiraling closer together in increasingly destructive encounters before the final, catastrophic collision around 2021. The grazing impacts would produce some debris but relatively little infrared energy. The full collision, by contrast, generated a debris field massive enough to create the chaotic dimming pattern that dominated the light curve from 2021 onward.

A Mirror for the Moon's Birth

What makes the Gaia20ehk discovery particularly arresting is not just that astronomers caught planets colliding in real time, but where in the system it happened.

The debris cloud appears to orbit Gaia20ehk at approximately one astronomical unit (AU) — the same distance that separates Earth from the Sun [1][2][3]. This orbital distance is significant because it places the collision squarely in the habitable zone of a sun-like star, and at the same location where, 4.5 billion years ago, a Mars-sized body called Theia is believed to have slammed into the young Earth.

That ancient cataclysm — known as the giant impact hypothesis — is the leading explanation for the origin of Earth's Moon [5]. The theory holds that Theia, an object with roughly 10% the mass of modern Earth, struck our planet at a glancing angle, ejecting a massive plume of vaporized rock and molten debris into orbit. That material coalesced, potentially within hours according to recent simulations, into the Moon [5][6].

"How rare is the event that created the Earth and moon? That question is fundamental to astrobiology," said Davenport [3]. If Moon-forming impacts are common, it would suggest that Earth-Moon type systems — with all their implications for tidal forces, axial stability, and the conditions that enabled complex life — may be widespread in the galaxy. If they are vanishingly rare, our planet may be even more exceptional than we thought.

The debris orbiting Gaia20ehk now offers a direct observational window into the earliest stages of this process. Over years, centuries, or millions of years, the scattered material could cool, clump together, and eventually solidify into new planetary bodies — potentially recreating something resembling an Earth-Moon system in another corner of the galaxy.

Source: GDELT Project

Data as of Mar 13, 2026CSV

A Growing Catalog of Cosmic Crashes

The Gaia20ehk event does not exist in isolation. Over the past several years, astronomers have assembled a small but striking portfolio of planetary collisions observed across the Milky Way, each revealing different facets of these violent encounters.

ASASSN-21qj: Ice Giants in the Night

In 2023, an international team published a landmark paper in Nature documenting the collision of two ice giant exoplanets around the star 2MASS J08152329-3859234, located approximately 1,850 light-years away in — coincidentally — the same constellation of Puppis [7][8].

The discovery had an improbable origin: a chance post on social media by an amateur researcher led professional astronomers to notice that the system had doubled in brightness at infrared wavelengths roughly three years before the star began to fade in visible light [8]. Computer models indicated the temperature, size, and duration of the glowing material were consistent with two ice giant planets — each several to tens of Earth masses — smashing together at a distance of 2 to 16 AU from the star. The resultant expanding debris cloud then traveled in front of the star approximately three years later, dimming it at visible wavelengths.

The ASASSN-21qj event demonstrated for the first time that collisions between large planets — not just rocky bodies or planetesimals — leave detectable afterglows that can persist for years.

Fomalhaut: The Disappearing Planet That Never Was

Perhaps the most dramatic revision in planetary collision science came from the Fomalhaut system, just 25 light-years from Earth. In 2008, the discovery of Fomalhaut b was hailed as the first directly imaged exoplanet in visible light — a triumph for observational astronomy [9][10].

But Fomalhaut b vanished. Follow-up observations failed to find it. In 2020, researchers proposed that what Hubble had photographed was not a planet at all, but an expanding cloud of dust produced by a collision between planetesimals — asteroid-like building blocks of planets [9]. The debris cloud, composed of particles roughly one-fiftieth the diameter of a human hair, had simply expanded and faded below detection limits.

Then, in December 2025, Hubble captured a second point source in the Fomalhaut system resembling the original appearance of Fomalhaut b twenty years earlier — evidence of yet another collision [10]. Theory had predicted one such collision every 100,000 years. Seeing two in just two decades suggests either the Fomalhaut system is unusually violent, or that planetesimal collisions are far more frequent than models predict.

HD 172555: Atmosphere Stripped by Impact

The young star HD 172555, a member of the Beta Pictoris moving group approximately 20 million years old, shows evidence of a different kind of planetary collision: one that stripped an Earth-sized world of its entire atmosphere roughly 200,000 years ago [11]. NASA's Spitzer Space Telescope detected signatures of vaporized and melted rock, along with SiO gas, pointing to a hypervelocity impact. More recently, JWST observations in 2025 detected emission from an inner gaseous disk containing neutral chlorine, sulfur, ionized nickel, and iron — the chemical fingerprints of a world torn apart [12].

The Detection Challenge

Observing planetary collisions has historically been compared to catching lightning in a bottle. These are brief events on astronomical timescales — the energetic aftermath may last years to decades, but in a galaxy where stars burn for billions of years, the window of detectability is vanishingly small.

The breakthroughs of the past few years owe much to the proliferation of automated sky surveys. The All Sky Automated Survey for Supernovae (ASAS-SN), which flagged the ASASSN-21qj event, scans the entire visible sky roughly every day [7]. The European Space Agency's Gaia mission, which alerted astronomers to the anomalous behavior of Gaia20ehk, has cataloged the positions and brightness of nearly two billion stars [1][2]. These surveys generate oceans of data that, combined with increasingly sophisticated machine learning and archival analysis techniques, allow researchers like Tzanidakis to spot needles in cosmic haystacks.

"It's incredible that various telescopes caught this impact in real time," Tzanidakis said. "There are only a few other planetary collisions of any kind on record" [4].

Source: NASA Exoplanet Archive

Data as of Mar 13, 2026CSV

Among more than 6,100 confirmed exoplanets discovered to date — the vast majority found via the transit method, which detects the tiny dip in starlight as a planet crosses in front of its star — planetary collision events represent an entirely different observational category [13]. Rather than the periodic, predictable dimming of a transit, collisions produce chaotic, evolving light curves that can confound automated detection pipelines. Finding them requires a combination of survey coverage, archival depth, and human intuition.

What Comes Next: The Rubin Observatory and a Hundred Collisions

The most transformative development on the horizon is the Vera C. Rubin Observatory, now operational in Chile with its 8.4-meter Simonyi Survey Telescope [14]. The observatory's Legacy Survey of Space and Time (LSST) will photograph the entire southern sky every few nights for a decade, creating an unprecedented time-lapse record of the dynamic universe.

The Gaia20ehk research team estimates that the Rubin Observatory could detect approximately 100 new planetary collisions over the next decade [3][4]. That projection, if borne out, would transform the study of planetary collisions from a field built on a handful of serendipitous discoveries into a statistically robust discipline — one capable of answering fundamental questions about how often worlds collide, what kinds of systems produce collisions, and how frequently those collisions produce the conditions for habitable planets.

"Andy's unique work leverages decades of data to find things happening slowly," said Davenport [2]. The Rubin Observatory promises to do the same at industrial scale, watching the sky with a patience and persistence that no human observer could match.

The Bigger Picture

Planetary collisions are not curiosities. They are, according to current models, an integral part of how planetary systems form and evolve. The rocky planets of our inner solar system — Mercury, Venus, Earth, and Mars — are thought to have reached their final sizes through a period of giant impacts lasting roughly 100 million years after the Sun's formation [5][6]. The Moon-forming impact was likely the last great collision in Earth's history, but the evidence from Gaia20ehk, ASASSN-21qj, Fomalhaut, and HD 172555 suggests that similar violence continues to reshape planetary systems across the galaxy.

Each observed collision opens a new window into the physics of planetary formation. The Gaia20ehk event, with its debris orbiting at one AU around a sun-like star, offers perhaps the closest analog yet to the conditions that gave rise to our own Moon. If the scattered material does eventually coalesce into a new satellite — a process that could take millions of years — it would provide direct confirmation of the giant impact hypothesis playing out in another star system.

For now, the scientific community is processing the implications of a discovery that Tzanidakis himself described with a word that has become the headline: completely bonkers. In the vast, quiet darkness between the stars, two worlds crashed together, and the light from that cataclysm traveled 11,000 years to tell us something profound about our own origins.