We have used the SINFONI near-infrared integral field unit on the VLT to resolve the optical emission line structure of one of the brightest (L~1e44 erg/s) and nearest (z=2.38) of all Lya blobs (LABs). The target, known in the literature as object 'B1' (Francis et al. 1996), lies at a redshift where the main optical emission lines are accessible in the observed near-infrared. We detect luminous [OIII]4959,5007A and Ha emission with a spatial extent of at least 32x40 kpc (4"x5"). The dominant optical emission line component shows relatively broad lines (600-800 km/s, FWHM) and line ratios consistent with AGN-photoionization. The new evidence for AGN photoionization, combined with previously detected CIV and luminous, warm infrared emission, suggest that B1 is the site of a hidden quasar. This is confirmed by the fact that [OII] is relatively weak compared to [OIII] (extinction-corrected [OIII]/[OII] of about 3.8), which is indicative of a high, Seyfert-like ionization parameter. From the [OIII] luminosity we infer a bolometric AGN luminosity of ~3e46 erg/s, and further conclude that the obscured AGN may be Compton-thick given existing X-ray limits. The large line widths observed are consistent with clouds moving within the narrow line region of a luminous QSO. The AGN scenario is capable of producing sufficient ionizing photons to power the Lya, even in the presence of dust. By performing a census of similar objects in the literature, we find that virtually all luminous LABs harbor obscured quasars. Based on simple duty-cycle arguments, we conclude that AGN are the main drivers of the Lya in LABs rather than the gravitational heating and subsequent cooling suggested by cold stream models. We also conclude that the empirical relation between LABs and overdense environments at high redshift must be due to a more fundamental correlation between AGN (or massive galaxies) and environment.

ApJ, Accepted (12 pages, 6 figures, 2 tables)