en.Wedoany.com Reported - Meta has launched the BoxCrete project, introducing artificial intelligence into the concrete design phase of infrastructure and construction. The platform leverages the Adaptive Experimentation (Ax) system, utilizing Bayesian optimization to explore the feasible space of material chemistry and physical formulations, thereby improving production efficiency, reducing costs, and enhancing ecological sustainability.

The technology core is trained on 123 mixture types and over 500 strength measurement data points, covering 69 mortar and 54 concrete types. The model can autonomously generate new formulations with a high probability of success and continuously optimize predictions through laboratory feedback, forming a closed-loop improvement cycle.

Meta notes that approximately 20% to 25% of the cement in the U.S. market relies on imports. BoxCrete can compare the performance of domestic and imported materials, helping local producers optimize resource use without compromising quality standards. This initiative directly serves the strategic goal of reducing foreign dependency.

Although AI cannot replace physical verification, on-site testing, and regulatory compliance, it can significantly shorten R&D cycles. By reducing trial-and-error iterations for variables like strength, construction convenience, and cost, the platform accelerates the implementation of sustainable solutions. BoxCrete has received the 2025 Construction Innovation Award, with partners including the spin-off company Amrize, the University of Illinois, and software provider Quadrel.

This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com