Global infrastructure systems are critically in need of upgrading to enhance their resilience against natural or man-made hazards exacerbated by climate change. Addressing this challenge while maintaining planetary boundaries requires a holistic and integrated approach to the design of the primary structural materials. Advanced Artificial Intelligence (AI) methods can aid the development of new sustainable material systems with superior performance in targeted applications. The current research advances an evidence-based formal and analytical framework enabling the design, optimization, and selection of sustainable structural materials in a data-driven manner. Consideration of embodied energy, recycling potential, and durability under extreme conditions over the whole life cycle is equally paramount. Existing trade-offs are explored for concrete and composite materials, where AI-based surrogate models are particularly well suited due to the abundance of local data. For structural alloys and lightweight metallic materials, the controlled synthesis of low-dimensional phase space clusters and Alloy Theory enable highly PID-oriented data generation for reduced-order, multi-objective surrogate-based design.