Machine Learning Finds 11,554 TESS Planet Candidates, Including 10,091 New Worlds
Updated
Updated · Earth.com · May 28
Machine Learning Finds 11,554 TESS Planet Candidates, Including 10,091 New Worlds
2 articles · Updated · Earth.com · May 28
11,554 planet candidates emerged from a machine-learning reanalysis of NASA TESS’s 2018 first-year archive, with 10,091 of them previously unreported.
84 million light curves from about 54 million stars were searched after Princeton-led researchers extended TESS data cleaning to fainter 16th-magnitude targets and trained separate classifiers for bright and dim stars.
One follow-up with the Magellan telescope in Chile already confirmed a hot Jupiter, while Joshua Roth said roughly 3,000 to 5,000 of the new candidates could survive vetting as real planets.
411 single-transit signals may point to longer-period worlds, and some candidates lie about 6,800 light-years away toward the Milky Way’s center—roughly doubling the survey volume TESS had effectively mined.
The catalog, published in The Astrophysical Journal Supplement Series, gives James Webb and future giant observatories a much larger target list and broadens planet-formation studies to fainter, more metal-poor stars.
With 10,000 new candidates, which worlds get priority from the Webb telescope, and how do we choose?
What secrets of planet formation are hidden around the dim, distant stars that AI has just unveiled?
How can astronomers be certain an AI's planetary discovery isn't just a clever algorithmic illusion?
Over 10,000 New Exoplanet Candidates Unveiled: How AI and TESS Are Transforming the Search for Habitable Worlds
Overview
In April 2026, a Princeton-led study made a landmark discovery by analyzing NASA’s TESS data and identifying over 11,000 exoplanet candidates, with more than 10,000 being entirely new. This breakthrough more than doubled the known TESS candidates and confirmed predictions that thousands of planets were hidden in the data. The achievement highlights the immense, previously untapped potential of TESS and marks the largest single increase in exoplanet candidates to date. As a result, our understanding of planetary systems beyond our own has expanded dramatically, opening new directions for exoplanet science and future exploration.