Various projects to find planets or entire planetary systems around main sequence stars in the solar neighborhood are presently under way. When ELTs will be operational, there will be literally thousands of confirmed planetary systems including spectro-photometric detections. At this point it becomes inevitable to consider the next logical step: the spectroscopic analysis of the atmospheres of these planets. High-resolution spectroscopy, i.e. resolving $v \times \sin (i)$ of these planets, in the wavelength regime of 950-5500nm is a powerful and promising tool. In view of the obvious contrast problems in detecting such planets non-LTE features are specifically targeted. Sensitivity estimates for the detection of the non-thermal OH glow in oxygen-bearing atmospheres are given. With 8m-class telescopes such a search is impossible, but a dedicated spectrograph, e.g. at the projected ESO 100m OWL telescope could detect Earth-like planets at a distance of ${\approx} 10$ parsec. A conceptual design for a dedicated spectrograph, NOCTUA, is presented. In case of ELTs of smaller size the science case changes and the instrument requirements have to be adjusted. Preparatory work with CRIRES, ESO's Cryogenic Infrared Echelle Spectrograph on the VLT at $\frac{\lambda}{\Delta \lambda} \approx 10^5$ as well as other science cases are shortly discussed.