Optimizing feather hydrolysate via machine learning for microbial recycling of waste concrete fines

Abstract BACKGROUND The concrete industry faces significant challenges from CO 2 emissions and the disposal of waste concrete fines (WCF). Microbially induced calcite precipitation (MICP) can bind WCF into bioconcrete, but the high cost of commercial culture media hinders its application. This study aimed to develop a low‐cost, sustainable medium from waste chicken feather hydrolysate, optimized via machine learning for two MICP bacteria, Sporosarcina pasteurii DSM 33 and Sutcliffiella cohnii DSM 6307. RESULTS The model successfully optimized the feather hydrolysis process based on bacterial growth by exploring parameters such as hydrolysis duration, temperature, and the concentration of feathers and hydroxide. The costs of the optimized hydrolysate were reduced by 80% compared to the commercial medium. The resulting bioconcrete samples demonstrated properties comparable to or superior to commercial media controls, including higher mechanical resistance and increased calcite content. The waste‐derived medium also promoted a beneficial shift in pore structure. Additionally, we found that mechanical stiffness is influenced not only by the CaCO 3 content, but also by its spatial distribution and the formation of crystal bridging, which proved to be critical. CONCLUSION Optimized feather hydrolysate is a viable substitute for commercial media, providing a sustainable, circular‐economy solution by valorising both poultry and concrete waste streams. © 2026 The Author(s). Journal of Chemical Technology and Biotechnology published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry (SCI).