On January 22, 2026, NASA shared a hopeful update for anyone who still looks up and wonders what’s out there. A refreshed, open-source AI software package called ExoMiner++ is now being used to dig into data from NASA’s Transiting Exoplanet Survey Satellite, better known as TESS.
Finding planets around other stars often starts with a tiny dip in brightness when a planet crosses in front of its star. The challenge is that space telescopes collect an enormous amount of data, and not every dip is a planet. Some are star activity, some are instrument noise, and some are other kinds of “false alarms.” NASA’s update is about making that sorting process more efficient, so researchers can focus their time on the best leads.
NASA notes that we’ve already discovered more than 6,000 exoplanets, with many found through NASA’s Kepler mission and TESS. But the story here is that the data still holds more secrets, and tools like ExoMiner++ are meant to help scientists comb through what we already have, more carefully and more quickly.
This is one of those quiet, steady wins that rarely looks dramatic in the moment. It’s a community tool, shared openly, aimed at turning a flood of raw measurements into clearer answers. Not every candidate will become a confirmed planet, but every improvement in how we search makes it more likely that the next “tiny dip” we notice really is a new world.
Source: https://science.nasa.gov/open-science/deep-learning-exoplanets-tess/