Tight binding calculations at the extended Huckel level have been used to study the bonding in NaCuTe, K2Cu5Te5, and CuTe, NaCuTe is predicted to be a semiconductor with the lowest band gap at Gamma. The highest filled orbitals are tellurium centered, and the lowest unoccupied orbitals are derived from Cu s. K2Cu5Te5 is metallic; the holes at the Fermi level are Te p,based (parallel to the a direction) and are delocalized equally over the three different Te sites. The reduction of K2Cu5Te5 by one electron is not expected to yield a semiconductor; a structural distortion is anticipated which will break Te-Te bonds. CuTe is predicted to be also metallic at the experimentally established geometry. An alternative structure where Te-Te bond pairing occurs has been optimized and found to be at a slightly lower energy than the experimental one. This pairing is predicted to open a band gap. However, experiment shows no evidence for a metal insulator transition down to 5 K. The reasons far why the extended Huckel method fails to correctly order these two structures in terms of energy are probed.