We present WSRT, VLA and VLBI observations of the Hi absorption in the radio galaxy NGC 315. The main result is that two Hi absorbing systems are detected against the central region. In addition to the known highly redshifted, very narrow component, we detect relatively broad (FWZI similar to 150 km s(-1)) absorption. This broad component is redshifted by similar to 80 km s(-1) compared to the systemic velocity, while the narrow absorption is redshifted similar to 490 km s(-1). Both Hi absorption components are spatially resolved at the pc-scale of the VLBI observations. The broad component shows strong gradients in density (or excitation) and velocity along the jet. We conclude that this gas is physically close to the AGN, although the nature of the gas resulting in the broad absorption is not completely clear. The possibility that it is entrained by the radio jet (and partly responsible of the deceleration of the jet) appears unlikely. Gas located in a thick circumnuclear toroidal structure, with orientation similar to the dusty, circumnuclear disk observed with HST, cannot be completely ruled out although it appears difficult to reconcile with the observed morphology and kinematics of the Hi. A perhaps more likely scenario is that the gas producing the broad absorption could be (directly or indirectly) connected with the fueling of the AGN, i.e. gas that is falling into the nucleus. If this is the case, the accretion rate derived is similar (considering all uncertainties) to that found for other X-ray luminous elliptical galaxies, although lower than that derived from the radio core luminosity for NGC 315. The data also show that, in contrast to the broad component, the density distribution of the narrow component is featureless. Moreover, in the WSRT observations we do detect a small amount of Hi in emission a few kpc SW of the AGN, coincident with faint optical absorption features and at velocities very similar to the narrow absorption. This suggests that the gas causing the narrow absorption is not close to the AGN and is more likely caused by clouds falling into NGC 315. The environment of NGC 315 turns out to be indeed quite gas rich since we detect five gas-rich companion galaxies in the immediate vicinity of NGC 315.