18:59, 12 November 2025

Russian Scientists Build an AI-Powered ‘Material Constructor’ for the Future

Researchers at Perm National Research Polytechnic University have developed an artificial intelligence system capable of designing entirely new materials — combining maximum strength with minimal weight for aerospace, medicine, and energy applications.

Engineering Matter in 3D

Scientists at Perm Polytechnic University have unveiled a next-generation AI-based material design system. This digital ‘constructor’ can independently generate microstructures with unprecedented combinations of properties — like extreme strength paired with ultra-low weight. Such capabilities open new horizons for the aerospace, medical, and energy industries.

At its core, the technology creates three-dimensional ‘digital blueprints’ of materials optimized for internal architecture. Unlike traditional approaches, which rely on massive computational resources or simply mix known solutions, this AI system can generate entirely new — yet manufacturable — material structures.

The breakthrough lies in the combination of two complementary approaches. First, a neural network trained on libraries of existing materials learns the fundamental principles of structure formation in 3D. Then, a genetic algorithm identifies the best combinations, evaluating each by parameters such as strength and density.

“Our neural network doesn’t just copy or blend known samples — it actually invents,” explains Evgeny Kononov, a research engineer at the university. “But what’s more important is that it does so systematically. All generated designs are automatically organized into a structured digital space, where similar microstructures are grouped nearby. This makes it easy to find, compare, and analyze different solutions — something that was practically impossible before.”

Finding the Perfect Balance

The key outcome of the system’s work is the discovery of so-called ‘limit solutions’ — the perfect compromise between conflicting requirements.

“At this stage, the system forms a set of structures in which both parameters cannot be improved simultaneously: if we increase strength, density inevitably rises, and vice versa,” explains Mikhail Tashkinov, head of the university’s research laboratory ‘Mechanics of Biocompatible Materials and Devices’. “Such outcomes are valuable because they define the best possible trade-offs between opposing demands.”

From Theory to Practice

Practical tests have confirmed the method’s effectiveness. The AI-generated 3D microstructures outperformed materials from the original database, demonstrating a 15–20% increase in stiffness at the same density.

The development has strategic significance not only for Russia but globally. It could dramatically reduce the time and cost of creating new materials for critical applications — from bone implants to lightweight aircraft components and heat-resistant turbine blades.

Industry and import substitution