A Jupiter-mass exomoon candidate circles a failed star, not a planet

A brown dwarf in the southern constellation Columba is rocking back and forth on a steady rhythm, and the best explanation is a body at least as heavy as Jupiter circling it. If the signal holds up, that companion is the first natural satellite ever detected with real confidence outside our Solar System, a prize astronomers have chased for decades without success.

The catch is that almost nothing about it looks like a moon as we know one. The object it circles is not a planet but a brown dwarf, a “failed star” too massive to be a planet and too small to burn hydrogen like the Sun, and that brown dwarf in turn orbits a small red star. The team behind the find, led by Kevin Hoy, a doctoral student at the European Southern Observatory in Chile, calls it an exosatellite and admits the vocabulary is struggling to keep up.

“This system is somewhat hard to define using Solar-System-based words like ‘planet’ and ‘moon’,” Hoy said in ESO’s announcement. “The exosatellite is clearly massive enough to be a planet, but it does not orbit a star, though it orbits an object that orbits a star. Being the third wheel in this system makes us want to call it a moon, even if it is nothing like the small, rocky moons we have in our system.”

The host star, CD-35 2722, lies about 73 light-years from Earth and has roughly half the Sun’s mass. Its brown dwarf companion, catalogued as CD-35 2722 B, weighs about 37 times as much as Jupiter. The newly reported satellite has a minimum mass of about 0.9 Jupiter masses and completes one lap around the brown dwarf roughly every 170 days.

How a wobble gave the hidden moon away

Nobody has seen the satellite. The team watched its gravity instead. As the companion swings around, it pulls the brown dwarf slightly toward Earth and then away, and those movements shift the lines in the brown dwarf’s spectrum by tiny, measurable amounts. This is the radial velocity method, the same technique that revealed 51 Pegasi b, the first planet found around a Sun-like star.

The measurements came from CRIRES+, a high-resolution infrared spectrograph on ESO’s Very Large Telescope at Paranal, in Chile’s Atacama Desert. Infrared is the right window here: brown dwarfs are cool and dim and give off most of their light at infrared wavelengths. And because CD-35 2722 B sits about 67 times farther from its star than Earth does from the Sun, the telescope can isolate the brown dwarf’s light rather than drowning in the star’s glare.

The astronomers gathered 26 radial velocity measurements over more than two years and tested several models against them, including one in which the brown dwarf hosts two satellites. A single companion on a 170-day orbit fit best. The strategy of hunting moons through the spectra of young, self-glowing giant objects was proposed years earlier in a paper co-written by Andrew Vanderburg, now an astronomer at Harvard University. “I never would have imagined that this technique would reveal such an unusual object!” Vanderburg told Eos, the news magazine of the American Geophysical Union.

Why no exomoon had been confirmed before

More than 6,000 planets have been found around other stars, yet their moons have stayed out of reach. Moons are small and faint, buried in the glare of both their planet and their star. Earlier candidates, such as a proposed Neptune-sized moon around the planet Kepler-1625b, came from tiny dips in starlight recorded by space telescopes, and follow-up analyses still disagree over whether that moon exists.

A separate team led by Quentin Kral recently used ESO’s Very Large Telescope Interferometer to find hints of a satellite in the HD 206893 system, but no firm detection. That record explains why the team words its own claim with care. “As exotic as it is, this system is truly unique and represents a breakthrough: the first plausible detection of an exosatellite,” said co-author Alice Zurlo, director of the Millennium Nucleus of Young Exoplanets and their Moons at Universidad Diego Portales.

A moon heavy enough to be a planet

Scale is where the system turns strange. Ganymede, the largest moon in the Solar System, is bigger than the planet Mercury but holds less than one ten-thousandth of Jupiter’s mass. On the published figures, the candidate around CD-35 2722 B carries at least about 2 percent of its host’s mass, roughly double the ratio between our Moon and Earth, which is already unusually high by Solar System standards.

Zurlo described it as “a giant gaseous body orbiting a highly massive companion, itself several times the mass of Jupiter.” Circling a star on its own, an object this heavy would simply be called a gas giant planet. How it formed is an open question. Our Moon most likely coalesced from debris after a giant impact on the young Earth, and Jupiter’s large moons grew in a disk of gas and dust around the young planet, but neither picture has been tested for a satellite this massive.

The system’s youth helps make the measurement possible. CD-35 2722 belongs to the AB Doradus moving group, a family of stars roughly 100 million years old, so the brown dwarf is still warm from its formation and glows brightly enough in infrared for precise spectra.

What the signal doesn’t settle

The detection rests on a repeating signal in one set of measurements, and the radial velocity method gives only a minimum mass. Because the tilt of the orbit relative to our line of sight is unknown, the true mass could be higher. The satellite has not been imaged, and it has not been caught passing in front of the brown dwarf.

There is also a possible explanation that involves no satellite at all. Something happening on or inside the brown dwarf, such as spots or weather rotating in and out of view, could in principle mimic a regular shift in its spectrum. Hoy told Eos the team “can’t think of any physical mechanism that could reproduce the signal we see,” which is why they consider a satellite the most likely explanation. That is a strong statement, but it is not the same as ruling every alternative out, and the paper presents the result as evidence rather than confirmation.

Then there is the label. There is no officially accepted definition of an exomoon, and the authors note it is uncertain whether this object will meet whatever criteria eventually emerge. “Perhaps we are approaching the limit of language invented to describe the Solar System, which is entirely unlike CD-35 2722,” the paper reads.

Common questions about the first exomoon

What is an exomoon?

An exomoon is a natural satellite orbiting a planet or another object outside our Solar System. There is no officially accepted definition, which is why the team behind the CD-35 2722 find uses the broader term exosatellite.

Has an exomoon ever been confirmed?

Not with certainty. Earlier candidates, including one around Kepler-1625b, remain disputed. The object orbiting the brown dwarf CD-35 2722 B is described by its discoverers as the first plausible detection of an exosatellite, and it still awaits independent confirmation.

What is a brown dwarf?

A brown dwarf is an object heavier than a planet but too light to sustain hydrogen fusion in its core like a true star. Most weigh between roughly 13 and 80 times the mass of Jupiter. They glow faintly in infrared light, especially when young, and slowly cool over billions of years.

How can astronomers detect a moon they cannot see?

Through its gravity. The satellite’s pull makes the brown dwarf move slightly toward and away from Earth, which shifts the lines in its spectrum. Measuring those shifts over many months reveals the length of the orbit and a minimum mass for the hidden companion.

How far away is CD-35 2722?

The star lies about 73 light-years from Earth in the southern constellation Columba. It is a small red dwarf, far too dim to see without a telescope.

What comes next

The study appeared in Nature on July 22, 2026, and drew a new wave of coverage this month. It draws on 26 spectra collected with CRIRES+ between October 2023 and February 2026, and its method builds on the moon-hunting strategy Vanderburg and colleagues proposed in 2018. The brown dwarf itself was discovered in direct images from the Gemini NICI Planet-Finding Campaign, reported in 2011. The next leap depends on ESO’s Extremely Large Telescope, under construction in the Atacama Desert: with its 39-metre main mirror, ESO says it will let astronomers detect smaller exomoons, the kind that would look far more like the moons of our own Solar System.

Reference: Hoy et al., “Planetary-mass exosatellite detected around the substellar companion of a star,” Nature, 2026. DOI: 10.1038/s41586-026-10751-w

Tags: astronomy, ESO, Very Large Telescope, brown dwarf, CD-35 2722, exomoon