This study explores the energy-absorbing efficiency of additively manufactured polymer structures using a self-driving lab (SDL) to perform over 25,000 physical experiments on generalized cylindrical shells. The experiments are selected and executed collaboratively by a human team and the SDL, which uses Bayesian optimization to choose the next design to test. The result is a structure with a 75.2% energy-absorbing efficiency, surpassing the previous record held by nature. The campaign also yields a large dataset that provides insights into the design principles for tough structures, including the importance of material stiffness, plasticity, and relative density. The findings highlight the potential of combining computational and experimental approaches to advance the design of energy-absorbing materials.This study explores the energy-absorbing efficiency of additively manufactured polymer structures using a self-driving lab (SDL) to perform over 25,000 physical experiments on generalized cylindrical shells. The experiments are selected and executed collaboratively by a human team and the SDL, which uses Bayesian optimization to choose the next design to test. The result is a structure with a 75.2% energy-absorbing efficiency, surpassing the previous record held by nature. The campaign also yields a large dataset that provides insights into the design principles for tough structures, including the importance of material stiffness, plasticity, and relative density. The findings highlight the potential of combining computational and experimental approaches to advance the design of energy-absorbing materials.