According to NASA’s December 2025 announcement that construction of the Nancy Grace Roman Space Telescope had been completed, the observatory is on track for launch by May 2027, with a possible earlier launch in fall 2026. The five-year primary mission is expected to produce one of the largest exoplanet catalogues in the history of astronomy. Julie McEnery, Roman’s senior project scientist at NASA Goddard, summarised the projections in the announcement: “In the mission’s first five years, it’s expected to unveil more than 100,000 distant worlds, hundreds of millions of stars, and billions of galaxies.”
The 100,000 figure refers specifically to transiting planets, the kind that produce a small dip in their host star’s brightness when they pass between the star and the telescope. That is the same technique NASA’s Kepler mission used to find roughly 2,800 confirmed exoplanets in the 2010s. Roman is built to do the same thing on a much larger scale. The wider survey area, the higher sensitivity in the infrared, and the deeper view into the dense centre of the Milky Way are what bring the number up by close to two orders of magnitude.
What is genuinely new is the second technique Roman will use in parallel, and the kinds of objects it will find with it.
The technique that finds planets without stars
Roman’s other planet-finding method is gravitational microlensing, a phenomenon predicted by Albert Einstein in the 1930s and used in practical astronomy since the late 1980s. Microlensing exploits the fact that mass bends light. When a foreground object passes nearly in front of a background star from the observer’s perspective, the gravity of the foreground object focuses some of the background star’s light toward the observer, producing a brief brightening that lasts hours to weeks depending on the mass involved.
Microlensing has two advantages over the transit method. The first is that it works regardless of whether the planet ever transits its star from our line of sight. The second, more important advantage is that it does not require the planet to have a star at all. A planet drifting alone through interstellar space will produce a microlensing signature as it passes in front of a distant background star. This is the only currently practical way to detect such objects.
Planets without host stars are called rogue planets, or free-floating planets in the more technical literature. They are not as exotic as the name suggests. Theoretical models have predicted their existence for decades, and the first candidates were detected by ground-based microlensing surveys in the 1990s and 2000s. The OGLE and MOA collaborations have steadily added candidates to the catalogue since. The James Webb Space Telescope’s 2023 observations of the Trapezium Cluster in the Orion Nebula added several hundred more candidates, including some in apparent gravitationally bound pairs that have not yet been fully explained. Roman is not the telescope that will find the first rogue planets. It is the telescope that will turn the catalogue from “a few dozen confirmed cases” into something approaching a population census.
How many, and of what kind
The current best estimate of what Roman will see in microlensing comes from a 2023 paper by Naoki Koshimoto of Osaka University and collaborators at NASA Goddard. According to NASA’s account of the study, Koshimoto’s analysis of existing microlensing data, combined with simulations of Roman’s expected performance, predicts that the telescope will detect approximately 400 Earth-mass rogue planets during its primary mission. That figure is roughly eight times the previous best estimate of 50, which had been based on the assumption that rogue planets are about as common as star-bound planets. The newer estimate, drawing on a candidate terrestrial-mass rogue planet already identified in ground-based data, implies that low-mass rogue planets are far more abundant than that.
Roman’s microlensing programme will also reveal rogue worlds at higher masses, including roughly Jupiter-mass and Neptune-mass free-floaters, and at lower masses than ground-based observations have been able to reach. As Koshimoto put it in the NASA materials, “Roman will be sensitive to even lower-mass rogue planets since it will observe from space. The combination of Roman’s wide view and sharp vision will allow us to study the objects it finds in more detail than we can do using only ground-based telescopes.”
The microlensing programme will also find roughly 1,000 ordinary planets orbiting host stars, including planets in orbits comparable to those of every planet in our solar system except Mercury. These are the planets that transit surveys miss because their orbital periods are too long for a transit to be reliably observed in a finite survey, and the only way to find them in any numbers is the kind of long-baseline microlensing campaign Roman has been designed to conduct.
The Galactic Bulge Time-Domain Survey
The microlensing work will be done by what NASA calls the Galactic Bulge Time-Domain Survey, one of three core surveys Roman will conduct during its five-year primary mission. The survey will repeatedly observe six fields near the centre of the Milky Way, watching hundreds of millions of stars at once for the brief brightening events that signal microlensing. The galactic bulge is the right place to look because it has the highest density of background stars to lens, and the highest probability of foreground objects, including planets, passing in front of them.
According to NASA’s technical description of the survey, the GBTDS is built around six 72-day high-cadence seasons, three early in the mission and three late, during which Roman will image each of the six bulge fields every 12.1 minutes in its wide F146 filter. Between the high-cadence seasons, four low-cadence observing periods spread across the middle of the mission will return to the same fields at intervals of a few days, allowing the survey to catch microlensing events that play out over longer timescales, including those signalling isolated stellar-mass black holes. Roughly 438 days of total observing time are budgeted across the five years. The combination of short cadence within seasons and long baseline across the mission is what gives the survey its sensitivity range, from terrestrial-mass rogue planets that produce microlensing events lasting hours to multi-year events from much heavier objects.
Why this matters
The current confirmed exoplanet count is roughly 6,000, accumulated over thirty years and across multiple ground- and space-based observatories. Roman is expected to multiply that number by roughly seventeen in five years, and to add the first robust statistical sample of rogue worlds. The point of finding so many is not primarily to add to the list. It is to allow population-level statistics: how common are Earth-sized planets in habitable zones, what is the mass function of free-floating planets, how does the planet population vary with location in the galaxy. Each of these questions is currently answered with a few dozen objects. Roman will provide enough statistical power to answer them with confidence.
Roman carries two instruments. The Wide Field Instrument, with its 0.28-square-degree field of view, will do the survey work that produces the 100,000 transiting planets and the rogue-planet catalogue. The Coronagraph Instrument, a smaller technology-demonstration instrument, will directly image a handful of large planets around nearby stars, testing techniques intended for future missions designed to image Earth-like worlds around Sun-like stars. The launch is scheduled for no later than May 2027, with the team currently aiming for a possible earlier window in fall 2026. The observatory will travel to the second Earth-Sun Lagrange point, about a million miles from Earth, the same orbital station occupied by the JWST. Whether the 100,000 figure proves precisely right is something that will be known only after the data arrives. What is less in doubt is that Roman is the first telescope designed from the start to do exoplanet science at this scale, and the first one with a serious prospect of finding rogue planets in numbers large enough to characterise them as a population rather than as individual curiosities.
