There is a planet 850 light-years away where sunset is hotter than sunrise. Not by a little. By enough that water molecules, caught in the evening breeze, are torn apart before they can fall.
WASP-121 b is one of those “ultrahot Jupiters” — gas giants skimming so close to their stars that one hemisphere is locked in permanent noon, roasting at 2,500°C, while the other faces an eternal midnight barely touching 725°C. The planet spins exactly once per orbit, like the Moon around Earth, so there is no rotation to mix the air. You would expect the boundary between day and night — the terminator — to be more or less symmetrical. Dawn and dusk should be equally miserable.
They are not. JWST observations published this week show the evening side absorbs measurably more infrared light than the morning side. The trailing edge, where the planet is turning away from its star, is hotter and more expanded than the leading edge where it turns toward the light. The evening terminator blocks more starlight. The water signal drops off there — not because the atmosphere is thinner, but because the temperature is high enough to break H₂O into its constituent atoms.
The culprit is wind. Ferocious eastward jets carry heat from the dayside into the night, but they do so asymmetrically. Because the planet rotates in the same direction it orbits, the wind piles heat onto the evening terminator — the dusk zone — while the morning terminator is relatively shielded. The result is a world where the sunset is the hottest part of the transition, not the sunrise. It is a small detail with large implications: our atmospheric models of these planets had assumed the terminators were roughly identical. They are not. The data even show a larger amplitude than the best models predict, which suggests we are missing something — probably clouds made of minerals like silicates, not water droplets, which current models mostly ignore because simulating cloud physics in a 2,500°C wind is computationally brutal.
What strikes me about this finding is not the temperature itself. It is the asymmetry. Physics rarely insists on perfect symmetry, but we build our models as if it should. We assume dawn and dusk are mirror images because it is easier to calculate. The planet does not care. It simply arranges itself according to the wind, the rotation, the tidal lock, and the chemistry of water breaking apart. The evening is hotter because the physics demands it, not because anyone predicted it elegantly.
There is something quietly humbling about watching a telescope catch a planet in the act of being more complicated than our equations. We have known WASP-121 b existed since 2015. We have measured its temperature, its orbit, its mass. But it took JWST’s NIRSpec instrument — watching the planet transit its star with enough precision to map longitude by longitude — to notice that the evening side is a distinct place from the morning side. Time resolved into space. The transit’s duration became the planet’s circumference.
I keep thinking about what it would mean to stand on that evening terminator, if standing were possible. The wind would be carrying the day’s heat into the night, and the sky would be full of silicate clouds — stone suspended in air, catching the last light. The water would already be gone, split into hydrogen and oxygen before it could condense. There would be no rain, no fog, no gentle cooling. Just the hot breath of the day being pushed over the edge into darkness, and the planet slowly expanding as it does, as if exhaling.
We found this by measuring the light blocked by a world we will never visit. A 30-degree rotation during transit, a slight dip in brightness at the end, a carbon monoxide signal that rises not because there is more CO but because the whole atmosphere is hotter. Small deviations, patiently observed, and suddenly the model is wrong and the planet is richer.
The researchers are already looking for other targets. There are more ultrahot Jupiters with the right geometry. Each one will have its own wind, its own clouds, its own preference for dawn or dusk. We are building a taxonomy of alien sunsets, and the first lesson is that no two terminators are the same.
That feels like the right kind of discovery. Not the answer to a question, but the beginning of a better question. What else are we assuming is symmetrical that is not? What else have our models smoothed over for convenience? The evening side of WASP-121 b was always hotter than the morning. It was waiting for us to look long enough to notice.
Sources: Max Planck Society press release on JWST observations of WASP-121 b (June 2026); study led by Cyril Gapp (MPIA) and colleagues.