NASA Flies Surplus Spy Optics on the Roman Telescope

NASA is targeting 7:26 a.m. EDT Sunday to fly the $4.3 billion Nancy Grace Roman Space Telescope on a Falcon Heavy from Kennedy Space Center. The liftoff is nine months early. The 2.4-meter mirror at the core of the observatory started as unused spy-satellite hardware.

The National Reconnaissance Office offered NASA leftover telescope assemblies after a cancelled surveillance program. Those optics, built to stare at Earth, are now sealed for a ride to deep space, and they are forbidden from looking down.

NASA Got the Mirrors After a Spy Program Collapsed

In the 1990s the NRO poured money into Future Imagery Architecture, a next-generation spy-satellite effort that swelled after the September 11 attacks. Boeing ran late and over budget. Scientific American, drawing on a New York Times investigation, reported that the overrun reached $13 billion on a $5 billion plan. Officials killed the optical half in 2005, but hardware had already been built.

Those assemblies sat in a clean room in Rochester, New York, at Exelis, later bought by L3Harris. Michael Moore, then NASA’s acting deputy director for astrophysics, had old Air Force contacts and asked if the surplus might move. The first answer was no. By 2011 the answer had changed. An NRO spokesperson later said the agency had “identified surplus telescope assets that were no longer required” and that the assemblies “met or exceeded the specifications NASA required.”

Alan Dressler, now an emeritus astronomer at the Carnegie Institution for Science, walked into that Rochester room in 2012 with John Grunsfeld, then NASA’s top science official, and found a row of Hubble-class mirrors. “It was beautiful,” Grunsfeld said. Each primary was 2.4 meters across, the same aperture as Hubble, but the tubes were shorter and the field of view was wider. Engineers nicknamed them stubby Hubbles. NASA took the gift, then spent more than a decade stripping classified parts, adding cameras, and building a spacecraft around one of the two flight telescopes.

FROM SURVEILLANCE FLOP TO LAUNCH PAD

1. 2005: The NRO cancels the optical half of Future Imagery Architecture after delays and cost overruns, leaving finished telescope assemblies unused.
2. January 2011: The spy agency tells NASA the surplus optics exist. Moore, who had already asked once, gets a call that the position has changed.
3. June 2012: NASA announces it owns two 2.4-meter telescopes. Astronomers at Princeton spend the summer mapping them onto the planned Wide-Field Infrared Survey Telescope, then called WFIRST.
4. 2016: NASA formally starts WFIRST around one donated telescope, doubling the mirror size of the original 1.1- to 1.3-meter concept and making room for a coronagraph.
5. 2018-2020: Nancy Grace Roman dies at 93. NASA later names the mission for her, the agency’s first chief of astronomy and the so-called mother of Hubble. David Spergel, a Princeton cosmologist who advised the project, said it was cancelled five times in the president’s budget before Congress kept it alive.
6. August 2026: The finished observatory is fueled, sealed, and rolled to Launch Complex 39A, nine months ahead of the public schedule.

The optics were free. The rest was not. Scientific American put the final bill at about $4.3 billion for the spacecraft, instruments, launch, and five years of operations. Ars Technica reported that figure is about $300 million above the 2020 development estimate, which NASA blamed on the pandemic, and that project manager Jackie Townsend said the team still ran to a cost cap. Each donated telescope had been valued at least at $250 million when the gift was announced, according to later Space.com reporting, yet the conversion still cost a flagship budget.

One Roman Frame Matches 100 Hubble Shots

Hubble stares. Roman sweeps. NASA says the Wide Field Instrument is a 300-megapixel infrared camera with Hubble’s angular sharpness and a field of view at least 100 times larger. One Roman image covers a patch of sky about 1.5 times the apparent size of a full Moon. The focal plane uses 18 Teledyne detectors, each 4,096 by 4,096 pixels, sensitive from 0.48 to 2.3 micrometers.

That wide field is a leftover of the spy design. A reconnaissance telescope needs to rake ground quickly. A cosmology survey needs the same trick aimed the other way. Julie McEnery, Roman’s senior project scientist at NASA Goddard, has said a month of Roman time on the Milky Way would take Hubble about a century.

That one month of observations to survey our Milky Way galaxy would take about a century with Hubble.

Julie McEnery, Roman senior project scientist, NASA Goddard

The data will not sit behind a professor’s login. NASA says Roman’s images go public as soon as they are processed. McEnery’s line to NPR was that a high-school teacher in Kentucky can see the same files as a Princeton faculty member. A memory card on the spacecraft also carries more than 1.3 million names submitted this summer, riding along to L2 for no scientific reason except that people asked to go.

A second instrument, the Coronagraph, is a technology test. It is built to block starlight and try to photograph older, colder giant planets in closer orbits than the hot young worlds direct imaging usually finds. NASA casts that demo as a step toward the Habitable Worlds Observatory concept, a future flagship meant to photograph Earth-like planets. Dmitry Savransky, a Cornell engineer on the coronagraph team, told his university that no other space observatory has this capability in visible light, and that a clean dataset would be unique if the test works.

What Roman Will Test About Dark Energy

Roman’s main science case is the same puzzle that created WFIRST. Type Ia supernovas in the late 1990s showed distant explosions fainter than a slowing universe should have produced. That result, which won the 2011 Nobel Prize in Physics, is the opening evidence for dark energy, the label for whatever is speeding up cosmic expansion.

The standard model that grew from that work is now under strain. Measurements of how fast the universe expands today disagree with the rate inferred from the early universe. Charles Bennett, a Johns Hopkins cosmologist, told Aerospace America the local figure is about 73 kilometers per second per megaparsec, plus or minus one, while the early-universe number sits around 68. He called the gap more than a decade old and growing more significant each year. Daniel Scolnic, a Duke University cosmologist, told NPR that recent data say the model might be wrong, and that Roman will pin down whether it is right or wrong.

THE HUBBLE TENSION FILE

• Local expansion: Bennett’s late-universe rate is about 73 km/s/Mpc, from supernovas and other nearby rungs on the distance ladder.
• Early-universe rate: The cosmic microwave background, read through the standard model, yields around 68 km/s/Mpc.
• Roman’s supernova haul: Survey designs described in Aerospace America project about 27,000 Type Ia supernovas, roughly ten times the samples now in hand, from a high-latitude time-domain campaign covering about 18 square degrees.
• Other checks: Weak lensing of galaxies, galaxy clustering, and time delays from strongly lensed supernovas give independent handles on expansion and on whether gravity itself needs a rewrite.

Jason Rhodes, a cosmologist at NASA’s Jet Propulsion Laboratory who works on Roman, told Scientific American that tensions between early- and late-universe measurements have some colleagues thinking cosmology is due for another shake-up. McEnery has said Roman can measure dark energy over larger distances and longer lookback times than the Vera C. Rubin Observatory or Europe’s Euclid mission, which concentrate on the nearby universe. Euclid will cover more sky with less detail; NASA plans to use Roman’s sharper overlapping fields to correct Euclid, then spread those corrections across Euclid’s wider maps. Rubin’s ground-based visible images can be paired with Roman’s infrared frames on the same patches.

Scientific American reported that in five years Roman is expected to survey about 12 percent of the sky and image billions of galaxies. Fuel is the likely limit. The primary mission is five years, with a design that supports five more, and NASA says the spacecraft can be refueled even though the agency has no servicing plan at L2 today.

Microlensing Reaches Planets Other Tools Miss

The other core survey stares at the galactic bulge and waits for stars to brighten when a foreground mass bends their light. Albert Einstein described the geometry in 1936 and doubted anyone would ever use it. Scott Gaudi, an astronomer at Ohio State University, helped turn that geometry into a planet hunt. He told NPR the method reaches worlds that are “just completely undetectable by any other method,” including planets toward the center of the Milky Way that resemble Jupiter, Saturn, Uranus, and Neptune.

Transit surveys need a planet to cross a star. Radial-velocity work needs a star to wobble. Microlensing needs only a chance alignment, so it is sensitive to cold planets far from their stars and to worlds with no star at all.

WORLDS OTHER METHODS MISS

• Bound cold planets: A 2019 yield study led by Matthew Penny forecast roughly 1,400 bound microlensing planets over a full five-year bulge survey, down to about a tenth of Earth’s mass under standard assumptions.
• Free-floating planets: A 2020 follow-up by Samson Johnson and colleagues, using the same fiducial mass function, estimated Roman will detect about 250 free-floating planets with masses down to that of Mars, including about 60 at Earth mass or below.
• A full census: NASA’s own mission pages put the microlensing goal above 1,000 exoplanets and describe a statistical census of planetary systems, including worlds at the galactic center that no transit pipeline can reach.

Those numbers move if the true population of homeless planets is higher or lower than the Cassan mass function the papers adopted. Johnson’s team said a log-uniform mass function could push free-floating detections toward 1,000. The point of the survey is to measure that function, not to assume it. Ground-based microlensing has already hinted that free-floating worlds are common; Roman’s cadence, measured in minutes rather than nights, is what lets a Mars-mass lens leave a blip instead of falling between observations.

NASA Still Has a Twin It Never Used

The NRO handed over two flight telescopes and spare glass for a third. NASA flew one. Scientific American reported that the agency later declined the extras because converting them still meant new instruments, a new spacecraft, and a launch, and that one remaining telescope had a slight fault in its mirror. NASA, the NRO, and L3Harris did not tell the magazine where the leftover hardware sits now.

The gift came with a hard rule. The telescopes may not be used to observe Earth. That condition, confirmed in SpaceNews reporting when NASA solicited other ideas in 2012, is the clearest mark of the hardware’s first life. A machine built to resolve objects on the ground smaller than a coffee mug, a capability Scientific American tied to a 2019 presidential tweet of an Iranian launch site, is allowed to map galaxies and forbidden to look home.

The mission almost died on paper anyway. Spergel’s five presidential-budget cancellations ran through the Obama years and the first Trump term. Thomas Zurbuchen, then NASA’s science chief, told Scientific American that putting Nancy Grace Roman’s name on the spacecraft was a way to make it harder to kill. Jared Isaacman, now the NASA administrator, said in April that when the agency gets a success like this, it should learn from the habits that produced it. Launching nine months early is the habit on display. The unused twin is the other half of the same story: a second Hubble-class mirror with no mission, no funding, and no permission to do the job it was built for.

The Falcon Heavy Window Opens Sunday Morning

On Aug. 21 teams sealed Roman inside the Falcon Heavy nose cone at Kennedy’s Payload Hazardous Servicing Facility. SpaceX said the observatory was encapsulated into Falcon Heavy’s fairing ahead of the Sunday attempt from pad 39A. A convoy moved the stacked fairing overnight on Aug. 24 and into the SpaceX hangar on Aug. 25. NASA’s Launch Services Program called this its fourth Falcon Heavy primary payload, after Psyche, GOES-U, and Europa Clipper. One side booster on this flight already flew on GOES-U.

NASA, the Roman team, and SpaceX scheduled a Launch Readiness Review on Friday, Aug. 28, for the last go/no-go poll. As of Friday afternoon, public accounts still listed Sunday. Live Science, citing National Weather Service forecasts on Thursday, put the chance of thunderstorms at Kennedy on Sunday at 50 percent and noted a backup window Monday at 7:22 a.m. EDT. August in Florida is the open variable. The hardware argument is over.

Coverage starts at 6:20 a.m. Sunday on NASA’s live channels. After separation, Roman heads for a halo orbit around Sun-Earth L2, about a million miles from Earth, the same gravitational parking neighborhood as Webb, with the two observatories kept far apart. SpaceX posted hangar footage of the three-core rocket waiting for that flight.

Falcon Heavy in the hangar at pad 39A in Florida ahead of Sunday’s launch of @NASA’s Nancy Grace Roman Space Telescope pic.twitter.com/Dvu5XnVxKX

— SpaceX (@SpaceX) August 26, 2026

Overnight crews have been living with that stack. Replies under the hangar post were not a debate about dark energy. They were people pointing at a rocket in a shed, including a parent whose son had worked through the night on the mating. The informed objection that keeps surfacing around the mission is older than the countdown: this telescope survived the budget office more than once, and the leftover spy program that supplied its mirror was itself a bust. Sunday’s risk is weather, not a new fight over the science case.

First Pictures Are Due in Early 2027

Even a clean launch does not mean pictures on Monday. NASA says the team will spend about three months on deployments, calibrations, and tests as the spacecraft cruises to L2 and checks out. Science operations start after that commissioning, and the agency expects to share the first images by early 2027.

Wendy Freedman, an astronomer at the University of Chicago, told NPR that when a new survey machine turns on, the surprise is often the finding nobody listed in the proposal. Roman’s listed jobs are already heavy: a verdict on dark energy, a census of cold and homeless planets, maps of how galaxies cluster under dark matter. The mirror doing that work was assembled for a different country, a different decade, and a different direction of gaze, and it is sitting on a Falcon Heavy because that other program failed first.