The gear that was built
The peer-reviewed Nature Materials experiment reads like a calculator: a microwave drives the antiferromagnetic resonance, the resonance jogs exciton susceptibility, and a homodyne bench reads coherent optical sidebands. Adak and colleagues built that chain in the layered antiferromagnet CrSBr, in a bulk crystal with no cavity boost, and saw conversion across about 300 MHz.[1]
Earlier magnon transducers leaned on a weak, off-resonant magneto-optical effect. The load-bearing part here is the exciton resonance. Several exciton-polariton resonances inherited the same magnon response, so the optical side is not stuck on one narrow line. That is which gear is doing the work.[1]
The part that is still off the bench
The authors write that higher cooperativity wants a smaller magnetic volume and a cavity. Those two parts are absent from this paper. A 300 MHz window in bulk does not measure the efficiency or the noise of a cavity device. There is also no comparison table; what you have is which interface actually ran.[1]
Another reading is available: the sideband could come from some other optical shift of the drive field, not from the magnon-exciton chain. The authors tied the homodyne readout to resonant excitonic susceptibility to push that reading back. An independent group publishing a cavity efficiency on the same crystal would tighten or loosen that reading.[1]
When the gear train would be complete
A useful transducer carries noise, bandwidth and chip integration together, before anyone counts qubits. This study showed the band; it did not bring a noise budget or a cavity efficiency. A later peer-reviewed measurement will either put a number on cooperativity in cavity CrSBr or show the bulk window shrinking on a chip.[1]