Where are the sources of the near infrared background ?

The observed near infrared background excess over light from known galaxies is commonly ascribed to redshifted radiation from early, very massive, PopIII stars. We show here that this interpretation must be discarded as it largely overpredicts the number of J-dropouts and Lyα emitters in ultra deep field searches. Independently of the detailed physics of Lyα line emission, J-dropouts limit the background excess fraction due to PopIII sources to be (at best) ≤ 1/24. As alternative explanations can either be rejected (e.g. miniquasars, decaying neutrinos) or appear unlikely (zodiacal light), whereas the reality of the excess is supported by the interpretation of the angular fluctuations, the origin of this component remains very puzzling. We briefly discuss possible hints to solve the problem.