A symmetry argument just split the 120-year-old magnetic sort

For 120 years, every magnetic material was sorted into one of two buckets, and the rule did not ask whether the buckets were complete. The 2026 Europhysics Prize to Libor Šmejkal, Jairo Sinova, and Tomas Jungwirth recognized that the rule, not the materials, was the part to rewrite. Altermagnetism is the third class: net magnetization that looks antiferromagnetic, ordering that is its own thing, and a switching behavior that lives in the gap.

Most readers will read this as a new-material story. The mechanism is a sorting story. The old taxonomy asked one question, does the net moment vanish, and stopped there. Altermagnets vanish in net moment and still have hidden order, so the classifier threw away the structure the device needs. TechTimes, citing the EPS prize citation, puts a number on what changed: more than 1,000 follow-on studies have filed into the bucket the old rule said was empty.

The repeatable move is to ask which hidden symmetry the existing two-bucket classifier is throwing away, and which industrial layer sits downstream. In spintronic memory the discarded axis required an external magnetic field to flip a bit. Altermagnets flip at terahertz without it, the constraint that bottlenecked STT-MRAM, the layer now in mass production on advanced-node microprocessors and the one flagged by a July 2026 Nature Physics review. The hardware gain is real because the classification is now defensible, not because anyone found a new rock.

The taxonomy precedes the supply chain. Whoever rewrites the sorting rule decides which axis the next product line gets to optimize.