In this paper the new concept of 'super-bridges', i.e. kilometre-long bridges suspended over carbon nanotube cables, is introduced. The analysis shows that the use of realistic (thus defective) carbon nanotube bundles as suspension cables can enlarge the current limit main span by a factor of ~3. Too large compliance and dynamic self-excited resonances could be avoided by additional strands, rendering the super-bridge anchored as a spider's cobweb. As an example, we have computed the limit main spans of the current existing 19 suspended-deck bridges longer than 1 km assuming them to have substituted their cables with carbon nanotube bundles (thus maintaining the same geometry, with the exception of the length) finding spans of up to ~6.3 km. We thus suggest that the design of the Messina bridge in Italy, which would require a main span of ~3.3 km, could benefit from the use of carbon nanotube bundles. We believe that their use represents a feasible and economically convenient solution. The plausibility of these affirmations is confirmed by a statistical analysis of the existing 100 longest suspended bridges, which follow a Zipf's law with an exponent of 1.1615: we have found a Moore-like (i.e. exponential) law, in which the doubling of the capacity (here the main span) per year is substituted by the factor 1.0138. Such a law predicts that the realization of the Messina bridge using conventional materials will only occur around the middle of the present century, whereas it could be expected in the near future if carbon nanotube bundles were used. A simple cost analysis concludes the paper.