The possibilities of breeding in liquid-metal fast breeder reactors (LMFBRs) and light water reactors (LWRs) are compared in two ways. The feasibility of breeding has been demonstrated in the Phénix reactor with a measured gain of 0.14. The gain in Superphénix will amount to ∼ 0.20. The studies show that while maintaining the performance of commercial reactors their breeding gain can be further increased either by the concept of heterogeneous cores or by using carbide or nitride fuel (breeding gain ∼ 0.35). Recently, the old idea of breeding in advanced pressurized water reactors (PWRs) has been taken up again with the objective of attaining a gain of 0.05. Unfortunately, these objectives had to be limited to a conversion ratio of 0.9 for safety reasons, and it is not certain whether operation will be rewarding economically. The strategy of substituting PWRs is examined using the French example. By gradually introducing LMFBRs, the cumulated uranium supplies in France can be kept within reasonable limits, which means that they attain three to jour times the home resources. This is not possible with advanced LWRs, which can be considered only as a possible backup solution for plutonium recycling into PWRs.