In a new study published by researchers at quantum computing company Quantinuum and collaborators from Caltech, Fermioniq, EPFL, and the Technical University of Munich, scientists have used Quantinuum’s powerful quantum computer, H2, to simulate a notoriously difficult system—quantum magnetism—in a way that pushes beyond what classical computers can reliably achieve.
“Digital quantum computers are much more flexible/universal, but we have paid for that flexibility with many technical challenges,” Dr. Michael Foss-Feig of Quantinuum and the paper’s lead author told the Debrief.
“This paper is an indication that we are finally moving these more flexible/universal machines into the realm of practical (and scientifically illuminating) quantum simulation,” Foss-Feig said.
Using A Common Framework
For decades, the question at the heart of the quantum computing conversation has been, “When will these machines be useful?” Based on the new research by Foss-Feig and is team, it seems an answer is finally taking shape.
The research team set their sights on simulating a well-known physics model: the quantum Ising model, which captures how tiny magnets—called “spins”—interact on a lattice. These spins prefer to align with their neighbors, but thanks to other effects, they can also “tunnel” between states, creating a rich playground for exploring how systems behave.
“What inspired us is that simulating quantum magnetism is in many ways the most natural type of physics simulation a quantum computer can do,” added Foss-Feig. “It’s widely agreed upon as the likely first place where quantum computers will start solving truly relevant problems in physics that can’t easily—or maybe just can’t—be solved by classical computers.”
From Theory to Quantum Experiment
Until now, demonstrations of so-called “quantum advantage” have mostly focused on problems designed to be hard for classical computers but offered little scientific utility.
Using Quantinuum’s System Model H2, the team ran simulations of the quantum Ising model that explored a phenomenon known as Floquet prethermalization. In this regime, a driven system temporarily mimics thermal equilibrium before eventually heating up. Understanding such dynamics is deeply important for fields ranging from condensed matter physics to device engineering.
This leap was made possible by the unprecedented fidelity of H2’s gates—particularly the two-qubit gates underpinning entanglement.
“Digitizing dynamics benefits from having two-qubit gates that impart small amounts of entanglement but have very high fidelity,” Foss-Feig noted. “…That’s why we were able to get clean results for deep enough quantum circuits to push beyond what can easily be simulated classically.”
Still, these simulations weren’t noise-free. The team developed sophisticated error mitigation techniques to clean up the data, allowing them to extract scientifically meaningful information about how the system evolved—including the emergence of hydrodynamic behavior and accurate estimates of physical quantities like diffusion constants.
Checking Their Work
To verify their quantum results, the team threw everything they had at the problem—tensor networks, neural networks, and operator compression methods—all among the best classical simulation techniques available today. But as Foss-Feig notes, these tools had limits.
“All of them could produce accurate results at short times, but eventually—well within the times we could reach on the quantum computer—they couldn’t keep up with the growth of complexity and entanglement intrinsic to the quantum dynamics we studied,” he says.
classified briefing
In other words, the quantum computer went further.
Foss-Feig is careful to acknowledge that with massive resources and enormous effort, some classical methods might still simulate the results achieved in this study.
“But we believe that’s only viable because we sit at the cusp of where quantum advantage kicks in—56 qubits,” he explained. “We don’t know of any classical methods that could keep up, even on the largest supercomputers, with a quantum computer like H2 if it had even 20 more qubits. Hence our excitement about Helios, our next system.”
Adding more Quantum Computing to the Equation
With H2 already challenging the limits of classical computing, the arrival of Helios, a 96-qubit system, later this year promises to blow past previous boundaries.
“My hope is that as digital quantum computers continue to improve and grow in scale, their ability to simulate a huge variety of quantum models will push quantum computing into the forefront of scientific computing very soon,” Foss-Feig said optimistically. “I already think we can positively impact the development of classical algorithms by offering a reliable benchmark, but I also think that in the very near future, quantum and classical methods will work side by side—especially in regimes where classical tools start to struggle.”
Kenna Hughes-Castleberry is a freelance science journalist and staff writer at The Debrief. Follow and connect with her on BlueSky or contact her via email at kenna@thedebrief.org