Astronomers have caught a disappearing planet in the act. TOI-2535 b is a Jupiter-sized world about 1,000 light-years away. For years it crossed in front of its star every 8.24 days, and in March 2026 it didn’t show up.
The planet is still there. Its orbit has tipped, so from Earth it now passes above or below its star instead of across it.
The finding is in a free-to-read paper in The Astrophysical Journal Letters, published Nov. 9. Samuel Yee of UCLA led it with 36 co-authors.
They call TOI-2535 b “the first confirmed planet orbiting a single star whose transits have been directly observed to disappear.” That qualifier matters, and so does the speed: the change played out within one telescope’s working life.
Why a planet can vanish without going anywhere
NASA’s TESS satellite, launched in 2018, hunts planets by the transit method. A planet that passes between its star and us blocks a sliver of starlight. The telescope records that as a small, regular dip in brightness.
That only works if we see the orbit almost edge-on. Tilt it a little and the planet’s path misses the star’s disk from our point of view.
Astronomers track this with the “impact parameter,” a measure of how far from the star’s center the planet’s path crosses. Zero is dead center. Near 1, the planet skims the edge, and above that it misses.
For TOI-2535 b, that number has moved fast. In 2019 the planet crossed near the middle, at about 0.2.
Later TESS visits put the track at about 0.6, then about 0.87. The paper puts the drift at 0.16 per year. That’s the fastest change in orbital tilt yet seen for a planet around a single star.
By our arithmetic, at 0.16 a year a track that starts near the center reaches the edge in about five years.
A track near the edge cuts a shorter line across the star, so each transit got shorter. The planet also covered less of the star, so the dips got shallower. When TESS looked again in March 2026, it found no transit within 12 hours of the predicted time.
The schedule was off by up to six hours
The tilt was not the only oddity. TOI-2535 b arrived early or late by as much as six hours, about 3% of its orbit. The authors say that’s the largest timing swing known for a hot Jupiter on an orbit shorter than 10 days.
“This is the most extreme example we’ve seen because the two planets are interacting so strongly,” Yee said in UCLA’s announcement. “The planet’s orbit is tilting so quickly that its transit disappeared while we were watching.”
Both effects point to a second planet. The team combined the transit data with 15 measurements of the star’s wobble.
These came from the NEID instrument on the WIYN 3.5-meter telescope at Kitt Peak, Arizona. The best fit is a nearby companion, probably at least as massive as Saturn. Its orbit sits at an angle of roughly 5 degrees or more to TOI-2535 b’s.
Two tilted orbits pull on each other, and both slowly swing around like the axis of a spinning top. Astronomers call this precession.
The team’s simulations give a cycle of about 30 to 50 years. The paper cites roughly 150 years for TOI-216 and 250 years for TOI-4504, two other precessing systems. A cycle this short is why seven years of data could capture the whole exit.
What the paper does not show
Nobody has seen the companion. The team inferred it from the timing and the star’s motion, and the paper says many different orbits fit.
The stable ones cluster near a 2:1 or 3:2 resonance, where the two planets’ orbital periods form a simple ratio. With only eight well-spaced timing points, the authors note, other solutions remain possible.
The 5-degree figure is a lower limit that rests on a modeling assumption. And the “first” needs its qualifier.
Planets that orbit pairs of stars often show fast transit changes, the paper notes. A Kepler candidate, KOI-120.01, also faded and vanished during that mission, though it may not be a planet at all.
UCLA says TOI-2535 b should transit again in 30 to 50 years. The paper treats that span as the length of the precession cycle and gives no date for the next transit. The hidden neighbor, UCLA adds, may cross the face of the star sooner.
A hot Jupiter with company is the bigger puzzle
Hot Jupiters are giant planets that circle their stars in days. Astronomers think most formed farther out and moved in.
UCLA’s release describes two routes. In the violent one, a near-collision flings the giant onto a stretched orbit, and the star’s gravity reels it in. In the gentle one, it drifts inward through the gas disk around a young star.
Astronomers know of hundreds of hot Jupiters, and most lack a massive close neighbor, the release says. “Because TOI-2535 b has a close neighbor, it probably migrated by slowly spiraling in instead of getting flung in,” Yee said.
There’s a catch. A gentle drift should leave a flat system, but this one has a tilt.
The paper suggests disk migration followed by something that later knocked the orbits out of line. It does not test that idea.
Measuring the star’s wobble is the next step
With the transits gone, TESS can no longer time the planet. “Since continued transit timing monitoring is not possible, a high-cadence RV campaign is the best path forward,” the authors write. RV means radial velocity, the star’s back-and-forth motion as its planets tug on it.
The models predict the companion moves the star by 30 to 120 meters per second. Instruments like NEID can measure that.
As with the K2-155 signal picked up by China’s FAST telescope, the data narrow the options without settling them yet. Orbits can shift within months closer to home, too: Earth’s temporary mini-moon, 2026 QQ8, leaves in December.
The next test is whether a run of NEID measurements pins down the companion’s mass and period. If it does, astronomers can predict when TOI-2535 b’s path will swing back across its star.