2026 QQ8: Earth’s New Mini-Moon Was Caught in April, Named in October and Leaves in December

Earth has had a second moon since April. Nobody saw it until August, and nobody could say what it was until 29 October. The rock is about a metre across, it is already more than 1.9 million kilometres away, and it leaves in December.

It is called 2026 QQ8, and for the first 71 days of its known life it did not have that name. Hungarian astronomer Krisztián Sárneczky found it on the night of 19 August at Piszkéstető Station in the Mátra Mountains, working with the station’s GINOP-KHK telescope. The first observation went into the record at 19.92213 Universal Time, just after midnight local time on 20 August, and the object went onto the Minor Planet Center’s confirmation page under a submitter’s tag, Sar2901. It stayed a tag for ten weeks.

The Minor Planet Center decides what counts as a rock and what counts as junk. On 2 September it declined to decide. Its editorial notice, MPEC 2026-R09, said the object had appeared on the confirmation page “with an orbit that is suggestive of either a mini-moon or an artificial object”, and that “the broader community seems to be split as to which is the most favored, with both opinions being proffered to the MPC”. A designation, the notice said, “would indicate we believe the object to be natural”, so the centre “would prefer to delay the assignment of a designation until the community is more broadly convinced that this is a natural object”. The rock had already faded from view, and the notice expected it to stay out of reach until mid-to-late November, when fresh observations might settle the question of the forces acting on its orbit.

On 29 October, at 16:50 Universal Time, the centre issued MPEC 2026-U95. It reads in full: “Orbit integrations indicate that this object is temporarily bound to the Earth. The orbital solution includes solar radiation pressure; the fitted effective area-to-mass ratio is (7.94 +/- 3.41) x 10^-4 m^2/kg, with the quoted uncertainty at 1 sigma. No link to a known artificial object has been found. Further observations and dynamical studies are strongly encouraged.” The next day the centre explained the reasoning: more observations had stretched the arc far enough to measure how hard sunlight pushes the rock, and the push matched a dense little stone, not a hollow piece of spacecraft. A natural object is heavy for its size. A spent rocket stage is mostly empty shell, and sunlight shoves it around.

Thirty-six observations fill the circular’s table, and their magnitudes fall the whole way down. On 19 and 20 August the rock shone at magnitude 20.3, and by 5 September it was magnitude 23.8, by 19 September 23.9. On 9 October, the last observation the circular lists, it was magnitude 24.77, picked up by David Tholen with the University of Hawaii’s 2.2-metre telescope on Maunakea. It had dimmed by about four and a half magnitudes in 51 days, and the circular counted 36 observations over that arc. The September notice had said the rock would not be observable again until mid-to-late November, but the big telescopes found it in September and October anyway. NASA’s Small-Body Database now lists 39 observations over an 80-day arc. The latest is dated 7 November, three days ago. The coverage published on 6 November that called 9 October the last observation is already out of date.

2026 QQ8 is the fourth rock Earth has caught this century. Each one was spotted months into its stay, each one was held back from a name while astronomers argued about junk, and each one left within months or years.

The first was 2006 RH120, a rock two to three metres across that rode with Earth from July 2006 to July 2007, never getting more than 1.74 million kilometres away. Eric Christensen found it on 14 September 2006 with the Catalina Sky Survey’s 27-inch Schmidt camera in Arizona, and it went onto the confirmation page as 6R10DB9. Then the centre took it off again, with the explanation that it “was not a minor planet”. Its spectrum matched the white titanium-oxide paint used on Saturn V rockets, so it might have been a piece of Apollo hardware after all. It took until 18 February 2008, seventeen months after discovery, for the object to get its designation. It comes back about every twenty years. Precovery observations from 11 September 2006, the ones that would have settled its orbit at the time, were only published on 31 July 2026, almost twenty years later.

2020 CD3 stayed the longest. Theodore Pruyne and Kacper Wierzchoś found it on 15 February 2020 at Mount Lemmon Observatory, a faint 20th-magnitude speck in Virgo, 280,000 kilometres away. The centre announced it ten days later. Its orbit says Earth caught it around 2016 or 2017 and let it go around 7 May 2020, a stay of about three years, and it was roughly a metre across. Its next close pass comes in March 2044, and the approach will most likely be too distant for another capture.

2024 PT5 barely counts. The ATLAS survey in South Africa found it on 7 August 2024, an 11-metre rock, and it was temporarily captured from 29 September to 25 November that year, 57 days, drifting just outside Earth’s Hill sphere. It never finished a single loop. That is the distinction behind the counting: some outlets call 2026 QQ8 only the third temporary moon ever found, because they count only the rocks that completed a full orbit. PT5 sits in the other column, the flybys. It comes back to Earth orbit in 2084, for about 43 days.

One of them was junk. 2020 SO turned up on 17 September 2020 in the Pan-STARRS survey in Hawaii, moving slowly enough and pushed around by sunlight enough to look artificial, and in December 2020 a spectrum from NASA’s Infrared Telescope Facility matched stainless steel. It was the Centaur upper stage from the Surveyor 2 launch of 20 September 1966, still circling the Sun 54 years later. The centre struck it from its database on 19 February 2021. That is why the centre waited 71 days on 2026 QQ8, and why the September notice reads the way it does: the last time a slow-moving object like this showed up, it was a rocket.

On Reddit, the readers heard the gap between the words. On the thread about the discovery, one commenter wrote, “Then I guess it’s not very captured is it?”, and another answered, “It’s being temporarily detained in orbit”. A third wrote, “We should send a ship or probe to grab it and bring it back! Finders keepers!”, and a fourth opened with the obvious joke: “That’s no moon…” The Minor Planet Center says the rock is “temporarily bound”. The press says Earth “captured” a moon. The jokes land because everyone can feel the first word doing more work than the second. Nobody can grab it. It is more than 1.9 million kilometres away, five times the distance to the Moon, and fading past magnitude 25.

Can anyone see it? No. It is a metre across and half as far away again as the Moon, and no amateur telescope will find it. The size is still argued over: the Minor Planet Center’s figure is one to two metres, the European Space Agency’s Near-Earth Object Coordination Centre puts it at 1.4 to three metres, and the coverage has called it everything from a washing machine to a dining table to an adult human. Its orbit takes 359.07 days, six days short of Earth’s year, on a path tilted half a degree from ours, and it comes past about once a decade. The agency’s close-approach table puts it about 1.35 million kilometres away on 13 December. Universe Magazine puts the escape around January 2027; most of the coverage says December. It will not come this close again until 2058. That pass is not expected to be close enough for another capture. Some astronomers think Earth may typically have up to half a dozen such objects riding along at any one time. We are only now getting good at spotting them. The imaging hardware keeps improving, and so does the software that reads the images. The public record for 2026 QQ8 runs to 39 observations over 80 days, the latest dated 7 November.