IBM boffins have emerged from their smoke-filled lab, claiming they have simulated real magnetic materials on a quantum computer, and that the output aligns with neutron-scattering results from physical experiments.
The work involved researchers affiliated with the US Department of Energy’s Quantum Science Centre, as well as a range of universities and national labs.
They picked KCuF3, a magnetic material that has been studied for years, making it a handy yardstick for spotting nonsense.
Classical machines struggle to model quantum systems without leaning hard on approximations that get wobbly as the maths explodes.
Quantum kit is meant to thrive there because it runs on the same odd rules, but the hardware has usually been too noisy to trust.
Biggish Blue claims that the difficult part has been solved.
It did not leave the quantum processor to do all the work, leaning instead on “quantum centric supercomputing” that splits tasks between quantum and classical systems.
A big part of the improvement came from error reduction, with IBM pointing to better two-qubit error rates as the unglamorous fix that made the numbers behave.
This is not a sign that quantum computers are ready to replace serious scientific computing any time soon, as KCuF3 is hardly the nastiest problem in materials science.
Still, getting a quantum simulation to track real experimental data is a sharper test than another abstract benchmark, and it gives the boffins something concrete to argue about.
If this approach holds up, it could feed into superconductors, battery research and drug development, but for now, the real fun is seeing what happens when IBM tries a messier material.
