The Next Exoplanet Problem: Thousands of Worlds We Can Find but Not Understand

More than 6,000 exoplanets are confirmed, and the census is expanding rapidly. But a confirmed planet is not necessarily an understood planet. Precise constraints on masses, interiors, and atmospheres are available only for a small, highly selected subset of targets. A new white paper submitted to ESO's call for future science questions argues that the next generation of planet-hunting missions will deepen this imbalance unless a new observational paradigm can characterize masses and atmospheres at population scale. The challenge of the 2040s, the authors write, is not discovery. It is interpretation at scale.

A census with a blind spot

Transit surveys like Kepler and TESS map large regions of parameter space in planetary radius and orbital period. That gives a planet's size and orbit. But robust constraints on planetary interiors and atmospheres require precise, jointly determined bulk properties—typically radius better than about 5% and mass better than about 15%—to break degeneracies among composition, structure, and evolutionary state. Mass constraints typically require high-precision radial-velocity measurements.

The paper notes that only about 10% of known giant planets meet those standards. For planets smaller than roughly four Earth radii, the figure falls below 1%. Consequently, the vast majority of known exoplanets lack the physical characterization needed to constrain their internal structures and atmospheric properties meaningfully. Most of the census remains physically ambiguous.

Why more detections widen the gap

That gap is set to widen. ESA's PLATO mission, the proposed Earth 2.0 observatory, and NASA's Nancy Grace Roman Space Telescope are expected to expand the census dramatically, including regimes poorly sampled by current surveys. They will densely populate the radius-period plane. Yet for the vast majority of detected planets, these surveys will provide little more than planetary radii and orbital periods; direct constraints on masses, interior structures, and atmospheres will remain available only for a small subset of favorable targets.

The limitation does not stem from insufficient instrumental precision, the paper argues, but from the prohibitive observing time required for large-scale follow-up. State-of-the-art high-resolution spectrographs on 5–8 m class telescopes can achieve the precision needed to measure the masses of even temperate Earth-sized planets orbiting solar-like stars. But reaching mass accuracies better than about 15% typically requires 800–1,200 radial velocity measurements per target. Such investments are incompatible with the multi-science nature and competitive time allocation of current and forthcoming facilities.

Atmospheric characterization faces a similar bottleneck. JWST has delivered unprecedented insights into exoplanet atmospheres, and the ELT with instruments like ANDES will enable transformative studies of individual planets, extending atmospheric characterization to smaller radii and, in favorable cases, to temperate rocky worlds. Ariel, ESA's planned large-scale atmospheric survey, will observe hundreds to thousands of planets, providing a homogeneous baseline. But its moderate spectral resolution and wavelength coverage limit its ability to resolve key degeneracies, particularly for small planets where stellar contamination and overlapping molecular features dominate. JWST and ELT-class facilities excel at detailed studies of individual systems but are intrinsically unsuited for large, homogeneous population surveys.

The photonic array proposal

The white paper's answer is a different kind of observatory. Instead of a single monolithic aperture, it outlines a photonically combined, fibre-fed modular array of telescopes feeding photonic lanterns, pupil remapping devices, and compact photonic spectrographs. The authors argue this could deliver an effective collecting area comparable to a 15–30 m telescope, with significantly lower cost per square metre and intrinsically stable, miniaturised, replicable spectrographs.

The required performance is demanding: broad spectral coverage from the visible to the near-infrared, about 0.5–1.8 micrometers; resolving power above 100,000; and long-term instrumental stability at the centimeter-per-second level. The paper links that stability to high-precision radial-velocity measurements and high-dispersion atmospheric spectroscopy. A scalable design could expand the array or replicate it at multiple sites, increasing survey speed and sky coverage.

The concept grows out of ongoing developments within the Photonic E-MARCOT consortium. It is a proposal, not a finished facility. The paper does not present new exoplanet measurements, nor does it claim the technical hurdles are solved. It does not present a cost estimate or an engineering timeline. The paper frames PLATO, Earth 2.0, Roman, and Ariel as the missions that will expand detection and therefore make population-level characterization urgent. The paper is a preprint, available as arXiv:2601.18841v1.

Still, the paper's central point is hard to escape. A list of planets is not the same as an understanding of planets. The 2040s challenge, the authors argue, is to jointly characterize masses, interiors, and atmospheres across statistically meaningful samples, so the census can be robustly linked to the physical processes of formation, evolution, and habitability.