Modular design allows several generations of products to co-exist in the installed base as product designs change to take advantage of improved performance via modular upgrades. Use of a common base platform and modular design approach allows a firm to offer updates for improved performance and flexibility via remanufacturing when products have multiple lifecycles. However, as the product evolves through multiple lifecycles, the large pool of product variants leads to the curse of product proliferation. In practice, product proliferation causes high levels of line congestion and results in longer lead times, higher inventory levels, and lower levels of customer service. To offer insights into the product proliferation problem, the authors employ a delayed differentiation model in a multiple lifecycle context. The delayed differentiation model gives flexibility to balance trade-offs between disassembly and reassembly costs by adaptively changing the push-pull boundary. An adaptive, evolving push-pull boundary provides flexibility for a remanufacturing firm to meet changing customer demands. The delayed differentiation model includes both a mixed-integer linear program and an analytical investigation of the evolutionary nature of the push-pull boundary. Both field observations and experimental results show that the nature of product proliferation and changing demand structures play significant roles in the cost and flexibility of the evolving delayed differentiation system.