Are you thinking of the adiabatic QC approach (i.e., the dwave machine)? I think they still haven't been able to show that it's faster than a classical approach for any problem.