Scientists at Aalto University and their international collaborators have used artificial intelligence to discover two new superconductors, YRu₃B₂ and LuRu₃B₂. The AI-driven approach identified the materials from a vast pool of candidates, and experimental tests confirmed they exhibit superconductivity. The research, published in Physical Review Research, demonstrates that machine learning can dramatically accelerate the search for materials that could one day lead to a room-temperature superconductor.
How machine learning speeds up materials discovery
Conventional methods for finding new superconductors rely on years of trial-and-error experiments and quantum-mechanical calculations. The new technique combines machine learning with theoretical physics to screen huge datasets of known compounds. The model predicts which combinations of elements are most likely to become superconductors, allowing researchers to focus only on the strongest candidates.
After the AI identified YRu₃B₂ and LuRu₃B₂ as promising, theoretical calculations backed the predictions. A team led by Professor Emilia Morosan at Rice University then synthesized the materials by combining the constituent elements. Laboratory tests confirmed superconductivity in both compounds, proving the AI-guided process works beyond theory.
Why room-temperature superconductors matter
Superconductors carry electric current with zero electrical resistance below a critical temperature. Most existing superconductors require extreme cooling near absolute zero (-273.15°C). A material that superconducts at room temperature would eliminate energy losses in power grids, make electric vehicles lighter and more efficient, and enable quantum computers to operate without complex cooling systems.
Professor Päivi Törmä, who leads the SuperC consortium, said the near-term goal is a practical room-temperature superconductor by 2033. "Superconductive materials that can operate at room temperature would forever change the way we consume energy. If such a material could replace regular conductors in applications like computers and data centres, global energy consumption could be slashed and the heat footprint of the ICT sector vastly reduced."
AI as a research partner
The discovery of YRu₃B₂ and LuRu₃B₂ is important not just for the materials themselves, but because it establishes a new method for scientific inquiry. The SuperC consortium, launched in 2023, brings together experts in quantum physics, materials science, and artificial intelligence to cut the time needed to identify new superconductors. For researchers in these fields, the ability to screen millions of candidates in a fraction of the time increases the odds of finding compounds that work at higher temperatures.
As AI becomes a standard tool in materials research, scientists can explore chemical spaces that were previously too vast to handle. The approach may soon be applied to other classes of materials, widening its impact across energy, electronics, and climate technologies.
Why this matters for science and research professionals
For researchers and scientists working in materials discovery, this study offers a concrete template for integrating machine learning into experimental workflows. The AI did not replace the human element-it narrowed the field so that experts could focus validation efforts where they count. Learning to apply similar methods can reduce the time from hypothesis to confirmed result. For those looking to build these skills, an AI Learning Path for Research Scientists provides structured training on using AI in scientific discovery. More broadly, the AI for Science & Research hub tracks developments where computational techniques meet laboratory science. Adopting these techniques now can position research teams to make faster, more efficient discoveries in the coming decade.
Your membership also unlocks:
