Roman opens the widest eye we have flown

NASA's Nancy Grace Roman Space Telescope lifted off from Launch Complex 39A at Kennedy Space Center at 7:26 a.m. EDT on 30 August 2026, aboard a SpaceX Falcon Heavy. The rocket released the observatory thirty-one minutes into flight; the side boosters returned to the launch site. Roman is now on a three-month cruise of roughly a million miles to its operating orbit, where commissioning begins.

The inherited mirror

Roman's 2.4-metre primary is the same diameter as Hubble's. It did not have to be made: it was donated to NASA, having originally been built for a reconnaissance satellite, and its existence is a large part of why a flagship-class observatory could be proposed at 4.3 billion dollars rather than considerably more.

The mirror is not what makes Roman interesting. The focal plane is. Behind the same aperture Hubble carries, Roman puts an infrared camera whose field of view covers around a hundred times the area of Hubble's in a single pointing, at comparable resolution. Roman does not see fainter or sharper. It sees more at once, and that difference converts a class of observation from impossible to routine.

Two dark problems

The elevator summary is that Roman studies dark energy, exoplanets and galaxy evolution. It is worth separating the first from the subject of this issue's lead story. Dark matter is a missing mass problem, attacked by underground detectors like LZ waiting for a particle to strike a nucleus. Dark energy is a missing-pressure problem: the accelerating expansion of the universe, which no detector can catch because it is not a particle. It is measured statistically, by mapping how structure grew over cosmic time.

Roman's method is to survey billions of galaxies and measure weak gravitational lensing — the percent-level distortions that intervening mass imprints on the shapes of background galaxies — alongside baryon acoustic oscillations and supernova distances. Each requires enormous samples and, more demandingly, exquisite control of systematic error. A survey telescope's accuracy is limited not by photon statistics but by how well the instrument understands itself.

The exoplanet programme runs on the same wide field by a different route: microlensing, the brief brightening that occurs when one star passes in front of another and its planet perturbs the light curve. The technique is uniquely sensitive to planets far from their stars and to free-floating worlds bound to no star at all — the population transit surveys structurally cannot see.

Nothing is expected from Roman for months. Commissioning a cryogenic infrared instrument is slow, and the first survey data will be released after that. But the instrument that will decide whether dark energy is a cosmological constant or something that changes with time is, as of 30 August, in flight.