A Hot Jupiter Vanished From TESS Because Its Orbit Tilted Out of View

Planets don't disappear. Their transits can, though, and astronomers have now watched one happen in real time.

TOI-2535 b is a hot Jupiter about 1,000 light-years away. It circles its star every 8.25 days. NASA's Transiting Exoplanet Survey Satellite (TESS) tracked it from April 2019 to March 2026, recording the small, regular dip in starlight each time the planet passed in front of its host. Then, in March 2026, the dips stopped. The planet is still there. Its orbit has tilted far enough that, from Earth, it no longer crosses the face of its star.

The result appears in a paper titled "A TESS Disappearing Act: The Vanishing Transits of TOI-2535 b," published in The Astrophysical Journal Letters and posted to arXiv on Nov. 9, 2026. The authors call it the first confirmed case of a planet orbiting a single star whose transits have been directly observed to disappear.

A transit is a lucky alignment

The transit method depends on geometry. We only see a planet cross its star if its orbit is lined up almost edge-on to our line of sight. As Universe Today's Evan Gough points out, that alignment is a matter of luck, and gravitational interactions inside a planetary system can make it temporary.

Astronomers describe how centrally a planet crosses its star with the impact parameter. A value near zero means the planet's path runs straight across the middle of the stellar disk. Larger values mean the path, or "transit chord," sits closer to the star's edge. Once the chord moves past the edge, there is nothing left to see.

Over seven years of TESS photometry, the team measured TOI-2535 b's impact parameter changing at a rate of 0.16 per year. Year by year, the planet's transit chord slid toward the limb of the star, and eventually moved off the stellar disk entirely.

Clocks running early and late

The tilting orbit was only one of the oddities. The transits were also badly off schedule. A planet on an undisturbed orbit should transit at intervals you can predict to the minute from a simple linear ephemeris. TOI-2535 b's mid-transit times strayed from that prediction by about six hours in either direction. According to Space.com's account of the work, transits arrived as much as six hours early or six hours late.

When a planet's transit timing drifts like that, something else is pulling on it. The researchers modeled transit-timing variations (TTVs) and transit-duration variations (TDVs) together, and added radial-velocity measurements from the NEID instrument, which track the star's back-and-forth wobble. Combined, the data point to a second planet that has not been seen directly.

Here is what the modeling says about that companion:

It is close to TOI-2535 b.
It is probably Saturn-mass or larger.
It is likely in mean-motion resonance with TOI-2535 b, meaning their orbital periods form a simple ratio, so the two planets keep tugging on each other at the same points in their orbits.
Its orbit is tilted relative to TOI-2535 b's, with a mutual inclination of roughly 5 degrees or more.

That mutual inclination is what matters. Two planets on misaligned orbits exert torques on each other that make their orbital planes precess, or slowly swing around. For TOI-2535 b, the swing has carried its orbit out of alignment with Earth.

"This is the most extreme example we've seen because the two planets are interacting so strongly," Samuel Yee, a UCLA postdoctoral fellow and the paper's corresponding author, said in a statement quoted by Space.com. "The planet's orbit is tilting so quickly that its transit disappeared while we were watching."

Two giants packed close together

The setup itself is unusual. The paper describes TOI-2535 as a system with two short-period giant planets. Hot Jupiters are rarely found with close planetary neighbors, which makes a massive, resonant, inclined companion near TOI-2535 b stand out. The authors call the system one of the most extreme examples of planet-planet interactions driving orbital evolution on human timescales.

Orbital changes like these usually play out over spans much longer than a career or a lifetime. In this case it took less than a decade of TESS monitoring to watch a transiting planet become a non-transiting one.

Why It Matters

TESS has been prolific. As of mid-October 2026 it had identified more than 8,100 candidate exoplanets, and just over 1,000 of them have been confirmed. Much of that catalog rests on the transit method, which quietly assumes that a planet transiting today will keep transiting.

TOI-2535 b shows that assumption can fail, and fail quickly. A planet's transit can be a temporary state, and in a dynamically active system the window can close within one survey's lifetime. That has two practical consequences.

First, follow-up planning. Astronomers who want to study a planet's atmosphere during transit, or refine its size, need the transits to continue. Strong timing variations in a hot Jupiter's transits could reasonably be read as a warning that the observing window may be limited.

Second, hidden companions. TOI-2535 b's neighbor was not found because it transits. It was inferred from the way it disturbs the planet we can see. Precise long-baseline timing, paired with radial-velocity data, can reveal planets that transit surveys miss on their own, and can constrain how their orbits are arranged relative to each other, including whether they are misaligned.

The system also offers a rare chance to watch gravitational dynamics directly instead of reconstructing them after the fact. The measured rate of change in the impact parameter, the size of the timing offsets, and the radial-velocity signal together constrain the unseen planet's mass and orbit in ways a single snapshot could not.

When will it come back?

Not soon. According to Space.com, TOI-2535 b is not expected to transit again for at least 30 years. The orbital plane will keep precessing, and at some point the transit chord should swing back across the star. Until then, the planet stays out of view for transit observers. The companion that caused this has not been seen directly either; so far it is known only from the way it pulls on TOI-2535 b.