Recent research has established that early-type galaxies in the local Universe and outside the Virgo cluster contain a significant amount of \ion{H}{i}. However, because of the long integration needed to detect this gas (typically found at low column-density) interferometric work in this field has so far been limited to small samples, and results have been hampered by poor statistics. This problem is made even worse by the large variety of \ion{H}{i} configurations in early-type galaxies: small clouds scattered around the optical body; tidal/accretion tails; on-going galaxy interactions; and more or less settled rotating discs and rings. Furthermore, the \ion{H}{i} phase does not seem to correlate with the stellar properties of the host galaxy (e.g., luminosity, age, kinematics), so that it is not clear what its role is in the evolution of these objects.

Within this framework, I will discuss the Atlas3D project, a multi-wavelength survey of a complete, volume-limited sample of 263 local early-type galaxies. Thanks to deep WSRT observations, Atlas3D represents a dramatic improvement in the study of the \ion{H}{i} in early-type galaxies at z=0. It enables a statistically meaningful analysis of, e.g., the relation between \ion{H}{i} and environment, the HI mass function, the dependence of \ion{H}{i} morphology/kinematics on host-galaxy properties. I will also discuss the possible extension (and its requirements and limitations) of this work to wider areas and higher redshift, which will naturally be enabled by the next generation radio instruments.