The youngest planet ever found was hiding in data nobody re-examined for 8 years

Under a million years old. Earth is about 4.5 billion. Every other exoplanet we have found was already finished, and Elias 2-24 b is not, which makes it the youngest exoplanet ever discovered by a wide margin.

It orbits a newly formed star roughly 450 light years away, somewhere between two and four Jupiter masses. It is still actively pulling material out of the disk of gas and dust it formed from. We are watching a giant planet eat.

The discovery route is the good part

In 2017 astronomers noticed a gap in that star's protoplanetary disk. Gaps are ring shaped clearings, the sort you expect if something has swept up material along its orbit, and a gap on its own is an inference rather than a sighting.

Confirmation came from archival Keck Observatory coronagraph data. A coronagraph blocks the star's glare so that light reflected off a planet can be picked out of what is left. The planet had been sitting in that data the entire time. Nobody had gone back and looked with the right question in mind.

It also settles an argument

The core accretion model says giant planets build up from the middle outwards, gathering material over time, rather than collapsing straight out of the disk in one go. One of its specific predictions is exactly this link between disk gaps and planets in the act of forming.

Elias 2-24 b is that prediction with a picture attached. It does not rule out disk instability as a formation route for other planets, but it puts a directly observed case on the accretion side of the ledger.

What to watch

Planet formation is the only stage of a planet's life we had never directly observed. We now have one caught mid-process, which turns a modelling argument into something you can point a telescope at again next year and measure the change.

The quieter point is about the archive. A meaningful fraction of what is currently discoverable is already sitting on a hard drive, waiting for someone to ask a question the original observers were not asking. That is true of Keck data from 2017 and it will be true of the Roman Space Telescope's output across its extended mission, which is being planned as a survey archive rather than a set of individual observations.

Watch the follow-up observations. A planet still accreting should measurably change, and with 3I/ATLAS we learned how much a second look at formation-era material is worth.