A New AI Sifts NASA Data and Confirms Over 100 Hidden Planets

Astronomers have unleashed a powerful new artificial-intelligence tool on a mountain of NASA data and come away with a treasure of new worlds. Reported on May 3, 2026, the system, named RAVEN (for RAnking and Validation of ExoplaNets), sifted through observations from NASA’s Transiting Exoplanet Survey Satellite, or TESS, and validated 118 brand-new exoplanets along with more than 2,000 additional promising candidates.

The work was led by Dr. Marina Lafarga Magro and colleagues at the University of Warwick, with the results published in the Monthly Notices of the Royal Astronomical Society. TESS detects planets by watching for the tiny dip in a star’s brightness when an orbiting world passes in front of it, but separating genuine planets from look-alike false alarms across millions of stars is painstaking work. RAVEN was trained to do that vetting consistently and at scale, freeing astronomers to focus on the most intriguing finds.

“Reported on May 3, 2026, the system, named RAVEN (for RAnking and Validation of ExoplaNets), sifted through observations from NASA’s Transiting Exoplanet Survey Satellite, or TESS, and validated 118 brand-new exoplanets along with more than 2,000 additional promising candidates.”

What makes the haul exciting is not just the numbers but the variety. The newly confirmed planets include rare ultra-short-period worlds that whip around their stars in under a day, and others nestled in the so-called Neptunian desert, a region where planets of a certain size are surprisingly scarce. By cataloguing so many objects in one consistent sweep, the team also sharpened estimates of how common different kinds of planets really are, in some cases reducing uncertainties by up to a factor of ten.

“This is not just a list of potential planets,” Dr. Lafarga Magro noted, but a sample reliable enough to map the prevalence of distinct types of worlds. The researchers are candid that each candidate still benefits from follow-up confirmation, and the tool complements rather than replaces human astronomers. Still, it is a hopeful glimpse of how thoughtful AI can turn vast archives gathering dust into fresh discoveries, helping us fill in the map of our galactic neighborhood far faster than before.