AI database points to rare-earth-free magnet candidates

Scientists at the University of New Hampshire have built the Northeast Materials Database, a searchable resource containing 67,573 magnetic compounds. The project used AI to read scientific papers, extract experimental data, and train computer models to identify magnetic materials and calculate the temperature at which they lose magnetism.

The database includes 25 materials not previously recognized as magnets capable of staying magnetic at high temperatures. Researchers say faster discovery of sustainable magnetic materials could reduce dependence on rare earth elements, cut costs for electric vehicles and renewable-energy systems, and support the U.S. manufacturing base.

Rare earth elements underpin many of today’s strongest magnets, which are used in smartphones, medical devices, power generators, electric vehicles, and other systems. The team said laboratory testing of every possible element combination could number in the millions, making AI-guided screening a faster route to candidates for future permanent magnets.