Chen, Hsien-ChunHsien-ChunChenSu, Zheng-ShunZheng-ShunSuChiang, YuanYuanChiangHui, Wei-HanWei-HanHuiChen, Tsai-ChinTsai-ChinChenGhimire, AshishAshishGhimireSHU-WEI CHANG2026-04-212026-04-212026-05-1513598368https://www.scopus.com/record/display.uri?eid=2-s2.0-105033240317&origin=resultslisthttps://scholars.lib.ntu.edu.tw/handle/123456789/737367Natural composites exhibit outstanding mechanical properties, refined over millions of years of adaptation through gradient architectures and hierarchical organization. Inspired by these principles, we present a strategy for designing hybrid bone–bamboo-like gradient composites and establish an integrated simulation-AI-3D-printing framework for their evaluation and validation. Using lattice spring simulation, we examine graded structures with monotonic, partially and randomly shuffled volume fraction sequences. Our results show that placing a high soft-phase volume fraction (46.80%) at the notch improves strength by ∼5% and toughness by ∼25%, due to enhanced crack deflection and distributed energy dissipation. To accelerate mechanical performance prediction, we train a deep learning model on 200 samples, achieving R2 scores of ∼0.8 for strength and toughness. The model generalizes across 1620 additional designs, capturing layer sequencing effects using only volume fraction inputs. Selected top-performing designs were 3D printed using a multimaterial printer, validating the simulation predictions. This work offers a mechanics-informed design framework that tunes notch-adjacent compliance through gradient sequencing to create hybrid gradient composites. Integrating natural design principles with AI and additive manufacturing, this study provides a pathway for developing tough bioinspired materials with tunable fracture resistance.false3D-printingBambooBioinspired compositesBone-inspiredGradientSimulationToughnessjournal article10.1016/j.compositesb.2026.1135952-s2.0-105033240317