Somewhere between your phone's vibration motor and the engine driving a Tesla, there's a dirty little secret the tech industry doesn't like to advertise: we're dangerously dependent on a handful of elements most people couldn't find on a periodic table, mined predominantly in one country that recently decided to squeeze the supply.
Welcome to the world of permanent magnets, where the most powerful stuff - neodymium-iron-boron, or NdFeB if you want to sound fancy at parties - has been the undisputed champion since the 1980s. These magnets are absurdly strong. They're in every electric vehicle, every wind turbine, every MRI machine, and probably the earbuds you're wearing right now. The problem? China controls about 90% of their production, and in April 2025, they started playing hardball with export controls. Ford literally had to halt production at its Chicago plant. Fun times.
The Hunt for Something Better (or at Least Different)
G. Jeffrey Snyder, a materials scientist at Northwestern University, just dropped a perspective piece in Advanced Materials arguing something provocative: we've barely scratched the surface of what's possible with magnetic materials. NdFeB magnets have complex crystal structures with, according to Snyder, millions of potential compositional variations we haven't even looked at yet. We found a good-enough solution in the '80s and basically stopped searching.
But "good enough" gets awkward when your supply chain depends on geopolitical goodwill.
The exciting part: we now have tools that didn't exist a decade ago. Machine learning models can predict magnetic properties before anyone touches a furnace. Autonomous laboratories can synthesize and test materials faster than any grad student (sorry, grad students). And high-throughput computational screening can sift through 32 million candidates looking for needles in a haystack the size of Montana.
What Would "Better" Even Look Like?
Snyder sets the bar high: magnets with saturation magnetization above 2.5 Tesla, or magnetic energy density exceeding 800 kJ/m³. For context, current top-tier NdFeB magnets sit around 1.4 Tesla and 500 kJ/m³. Finding something that blows past those numbers would be like discovering a battery with twice the energy density of lithium-ion - it would reshape entire industries.
Meanwhile, researchers aren't just dreaming. Georgetown University recently announced high-entropy borides - magnets made from iron, cobalt, nickel, manganese, and boron - that achieve strong magnetic anisotropy without any rare-earth elements. A team at the University of New Hampshire built an AI system that reads scientific papers, extracted data on 67,573 magnetic compounds, and identified 25 previously unknown materials that stay magnetic at high temperatures.
Why Your Next EV Might Care
Electric vehicles are magnetic hogs. A typical EV motor contains 1-2 kilograms of neodymium and praseodymium. With 94% of traction motors still using magnet-based technology and global demand approaching 385,000 tonnes annually, any disruption hits hard.
Some automakers are hedging their bets. BMW and Renault are switching to externally-excited synchronous motors that use copper coils instead of permanent magnets. GM invested in Niron Magnetics, which is developing rare-earth-free alternatives. But these are workarounds, not solutions. The physics still favors strong permanent magnets for power density.
The Database Revolution
One underappreciated bottleneck: we didn't have good data. Magnetic properties have been scattered across supplementary materials and inconsistent metadata for decades. The Northeast Materials Database (NEMAD), with its 67,573 entries, changes the game. Machine learning models trained on it can classify materials as ferromagnetic, antiferromagnetic, or non-magnetic with 90% accuracy.
This is the unsexy infrastructure work that makes breakthroughs possible - like building roads before you can have a car industry.
The Actual Possibility
Snyder's argument isn't that we'll definitely find super-magnets. It's that we haven't really tried with modern tools. High-throughput computational methods, combined with AI-driven hypothesis generation and robotic labs that never sleep, mean we can explore chemical space orders of magnitude faster than before.
The next great magnet might be hiding in a composition nobody thought to try because it seemed too weird. And with global supply chains looking increasingly fragile, "weird" is starting to sound pretty appealing.
References
-
Snyder, G. J. (2026). The Possibility of New Complex Magnet Materials. Advanced Materials, e18751. DOI: 10.1002/adma.202518751
-
Itani, S., Zhang, Y., & Zang, J. (2025). The northeast materials database for magnetic materials. Nature Communications. DOI: 10.1038/s41467-025-64458-z
-
Liu, K., Yin, G., & Beeson, W. (2025). High-entropy boride magnets using earth-abundant elements. Advanced Materials. Phys.org coverage
-
IDTechEx. (2025). 2025 to be a Defining Year for the Rare Earth Magnet Market. Research Article
-
Park, J. et al. (2025). Rare-Earth-Free Iron-Based Permanent Magnets: Progress, Challenges, and Perspectives. MetalMat. DOI: 10.1002/metm.70022
Disclaimer: This blog post is a simplified summary of published research for educational purposes. The accompanying illustration is artistic and does not depict actual model architectures, data, or experimental results. Always refer to the original paper for technical details.