The most serious challenge concerning genetically modified insects remains their invasion ability. Indeed, transgenic insects often show lower fitness than wild individuals, and the transgene does not seem able to spread through a natural population without a driving system. The use of remobilizable vectors, based on the invading properties of transposable elements, has been frequently suggested. Simulations show that this strategy can be efficient. Moreover, if the transgene is designed to use transposition machinery already present in the genome, the transgene invasion appears to be potentially reversible after a few hundred generations, leading to new experimental perspectives.