Some planets orbit so close to their stars that a single “year” lasts less than a day on Earth. A few of those worlds just turned up hiding in data NASA already had.
A team led by researchers at the University of Warwick has pulled 118 newly validated planets from archived observations collected by NASA’s Transiting Exoplanet Survey Satellite, better known as TESS. Thirty-one of those worlds had never been detected before. The results, published in Monthly Notices of the Royal Astronomical Society and highlighted by the university in May 2026, suggest that publicly available space telescope data still contain a trove of undiscovered planets waiting for sharper tools to find them.
How 118 planets emerged from old light curves
The discoveries came from a purpose-built tool called RAVEN (RAnking and Validation of ExoplaNets). RAVEN is a Bayesian statistical framework that combs through the tiny, repeating dips in starlight that signal a planet crossing in front of its host star. It scores each dip against a lineup of false-positive scenarios, such as eclipsing binary stars or background blends, then ranks candidates by their probability of being genuine planets.
The team ran RAVEN across full-frame image data from 55 TESS sectors, covering light curves for more than 2.2 million Sun-like (FGK-type) main-sequence stars processed through NASA’s Science Processing Operations Center pipeline. Beyond the 118 validated planets, the analysis flagged more than 2,000 additional high-probability candidates that scored well enough to justify follow-up observation but have not yet crossed the confirmation threshold.
Marina Lafarga Magro, Andreas Hadjigeorghiou, and Kaiming Cui are among the researchers named in the Warwick coverage. The team framed the work as proof of concept: reprocessing publicly available TESS data with modern automated ranking methods can still yield significant planet hauls, even from sectors that other search pipelines have already picked over. In that sense, RAVEN is less a one-off project than a blueprint for squeezing more science out of existing archives as algorithms improve.
Ultra-short-period worlds and the Neptunian desert
Among the confirmed planets are ultra-short-period worlds, rocky bodies that complete a full orbit in less than 24 hours. At those distances, a planet practically grazes its star. Surface temperatures can soar past 2,000 degrees Fahrenheit, hot enough to melt rock. Understanding how such worlds form, migrate inward, and survive without being torn apart by tidal forces remains one of the open puzzles in planetary science.
The study also turned up confirmed planets sitting inside the so-called Neptunian desert. In exoplanet research, that term describes a zone of orbital and size parameter space where Neptune-sized planets are conspicuously rare close to their stars. The leading explanation is that intense stellar radiation strips away the thick atmospheres of planets in that zone, shrinking gas-rich mini-Neptunes down to bare rocky cores. Finding confirmed examples in the desert gives theorists fresh data points for testing models of atmospheric loss and orbital migration.
For context, TESS has now contributed to the confirmation of thousands of exoplanets since its launch in 2018, but the mission’s full-frame images contain far more transit signals than any single pipeline has been able to vet. The RAVEN results underscore how much discovery potential remains locked inside data that is already sitting in public archives.
A companion study hints close-in planets are more common than thought
A related demographics study, currently available as an arXiv preprint, examines how often close-in planets appear around Sun-like stars in the TESS sample. Its preliminary conclusion: short-period planets may be more common than earlier surveys indicated, partly because many fall below current signal-to-noise detection limits.
That finding aligns with the large candidate pool RAVEN surfaced, but it carries an important caveat. The demographics paper has not yet completed peer review. Its occurrence-rate estimates could shift once referees scrutinize the detection-completeness models, especially for very small planets or systems around particularly noisy stars. For now, the “more common than expected” message is best treated as a promising trend rather than a settled conclusion.
The RAVEN authors also posted a preprint version of their own analysis, which broadly matches the peer-reviewed article and reinforces the main numerical claims.
What still needs to happen
Statistical validation is a powerful screening tool, but it is not the same as a direct mass measurement. The 31 newly detected planets have not yet been independently confirmed through ground-based radial-velocity observations or follow-up from other space telescopes. Some fraction could still turn out to be eclipsing binaries, blended background systems, or instrumental artifacts once subjected to high-resolution spectroscopy and imaging.
Those follow-up campaigns will also be essential for measuring planetary densities, the key to determining whether a given world is rocky, water-rich, or wrapped in a thick gas envelope. Facilities like the James Webb Space Telescope and ground-based spectrographs are well suited for that work, though telescope time is competitive and the queue is long.
There are also questions about how representative the RAVEN sample is. Because the tool focuses on Sun-like FGK stars within specific brightness ranges, it may under-sample planets around cooler M dwarf stars or hotter, more massive hosts. Ultra-short-period planets could form and evolve differently around low-mass stars, so extending the search to other stellar types will matter for any broader population claims.
Sector-by-sector consistency in the TESS data pipeline is another open question. The RAVEN team analyzed 55 sectors spanning several years of observations. Subtle variations in pointing stability, background noise, or calibration between early and late sectors could influence which candidates cross the validation threshold, potentially biasing the apparent distribution of the most extreme planets.
Why reprocessing old data keeps paying off
The practical takeaway for anyone tracking the exoplanet census is straightforward: TESS data already in public archives still contain undiscovered worlds, and the tools to find them are getting sharper. The more than 2,000 high-probability candidates identified in this work represent a pipeline of future confirmations that will keep researchers busy for years.
The next test will be whether re-running RAVEN on more recent TESS sectors, now armed with a training set that includes ultra-short-period and Neptunian desert planets, can push the search into even more challenging territory. Smaller, cooler worlds whose signals barely rise above the noise are the obvious frontier. If automated validation frameworks can reach those planets without sacrificing statistical rigor, the catalog of known worlds orbiting other stars is set to grow considerably, not from new missions, but from a closer look at the data we already have.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
