In February 2026, a team at UIUC took a thin film of yttrium iron garnet and punched hexagonal arrays of circular holes through it. Then they measured how spin waves — magnons, the collective precession of magnetic moments — propagated through what remained. What they found was not an analogy. Not an approximation. The spin waves obeyed the Dirac equation.
The same equation that governs electrons in graphene. The same linear dispersion, the same valley degeneracy, the same topological protection. Emerging from a completely different physical system — no quantum mechanics required, no carbon atoms, no atomic-scale precision. Just geometry: a honeycomb of absences in a magnetic continuum.
This is worth sitting with. The Dirac equation didn’t arrive because someone engineered it. It arrived because the lattice demanded it.
The structure is called an antidot lattice — not dots of material, but dots of absence. A hexagonal array of holes etched into a continuous magnetic film. Between the holes, narrow bridges of magnetic material remain. And on those bridges, spin waves get trapped.
The trapping creates localized orbital modes. Some are symmetric, s-like — round puddles of precession sitting at the honeycomb vertices. Others are antisymmetric, p-like — two-lobed oscillations oriented along the bridges. These modes couple through the bridges, hopping from site to site, forming bands. Nine of them.
Two of the nine bands come from s-orbitals on the honeycomb sublattice. They disperse, cross at the K point of the Brillouin zone, and form a cone. That cone is linear — energy proportional to momentum, zero effective mass. The Dirac cone. The same structure that gives graphene its extraordinary electron mobility, its Klein tunneling, its valley-polarized currents.
Above them, p-orbitals and kagome modes produce seven more bands with their own crossings and a flat band — the hallmark of frustrated geometry, where destructive interference pins waves in place. But the essential physics lives at that first cone: two sublattices, nearest-neighbor coupling, hexagonal symmetry. That’s all it takes.
The deep result is not that magnons can mimic electrons. It’s that the Dirac equation is not a fact about electrons. It’s a fact about honeycomb geometry with two-sublattice coupling. Carbon atoms produce it because they sit on a honeycomb. Holes in a magnetic film produce it because spin-wave modes sit on the same honeycomb. The substrate doesn’t matter. The lattice writes the equation.
And the equation carries its consequences. Topological protection means backscattering-immune edge states — spin wave channels that route around defects without loss. Valley selectivity means information encoded in which Dirac cone carries the wave. All of this at GHz frequencies, in micron-scale structures, controllable with magnetic fields and electric currents. Microscale circulators for 5G signal routing. Topological magnonic waveguides. The equation is portable, and it brings its physics with it.
At the Dirac frequency, spin waves propagate isotropically — circular wavefronts expanding at constant velocity through the lattice, as though the holes weren’t there. This is the signature of massless dispersion: no preferred direction, no slowing down. Move away from the Dirac frequency and the lattice reasserts itself. Wavefronts deform, hexagonal anisotropy emerges, group velocity becomes direction-dependent. The geometry that created the equation also defines the domain where it holds.
The same equation, written by geometry into different substrates. Perhaps this is what mathematics is — not a human invention, not a Platonic discovery, but the invariant that survives when you replace everything except the shape. The thing that remains when you change the medium but keep the lattice. Not the physics of electrons, or magnons, or any particular excitation. The physics of the honeycomb itself.