Watching Mass Form in Real Time: How Quantum Chips Are Simulating Particle Physics

Between early 2025 and mid-2026, quantum processors have gone from simulating toy models of particle physics to watching confinement, string breaking, and hadron scattering play out in real time. These are not full simulations of the strong force. They are the first hardware demonstrations that the phenomena behind nuclear matter can be reproduced on programmable quantum machines. Here is what was measured, what it means, and what still separates these results from the real thing.

The problem classical computers cannot solve

The Standard Model’s account of the strong nuclear force is quantum chromodynamics, or QCD. It describes how quarks exchange gluons inside protons and neutrons. It is spectacularly well-tested. The difficulty is computational. At low energies the force between quarks grows with distance, so perturbation theory fails. The only first-principles method is lattice gauge theory: discretize spacetime into a grid, place the fields on its links, and compute. Classical lattice QCD has delivered precise results for static properties: hadron masses, decay constants, form factors. What it cannot do is simulate real-time dynamics. The barrier is fundamental. Evolving a quantum field forward in time means tracking an exponentially large space of states. On a classical computer, the memory cost scales exponentially with system size. On a quantum computer, it scales linearly.

That linear scaling is why this field exists. Classical lattice QCD works by rotating to imaginary time. This converts quantum amplitudes into statistical weights a computer can sample. The trick breaks for real-time processes and for matter at finite density: the weights become complex, sampling fails, and the calculation collapses. This is the sign problem. Collisions, thermalisation, dense nuclear matter, the dynamics of the early universe: all sit on the wrong side of that barrier. A quantum computer does not face it.

Parameter	Classical lattice QCD	Quantum lattice gauge theory
Time evolution	Imaginary time (Wick rotated)	Real time
Resource scaling	Exponential memory cost	Linear qubit scaling
Primary bottleneck	Sign problem at real time and finite density	Gate fidelity, noise, qubit count
Best used for	Static properties: hadron masses, decay constants	Real-time dynamics: string breaking, collisions

Confinement, observed on a chip

Confinement is the reason you have never seen a free quark. Pull two quarks apart and the energy in the gluon field between them grows with distance. Eventually it becomes cheaper to create a new quark-antiquark pair from the QCD vacuum than to stretch the string further. This mechanism holds all nuclear matter together. A complete dynamical understanding of it remains beyond classical computation.

Superconducting In January 2025, Mildenberger and colleagues reported the first digital quantum simulation of confinement dynamics in Nature Physics. Working on a Google superconducting processor, the team simulated a Z₂ lattice gauge theory. This is a simplified model, but it captures confinement’s essential structure. Tuning a single parameter that controls electric-field energy, they drove the system from a deconfined phase into a confined one and watched the transition across 25 Trotter steps. The key achievement was efficiency: six native two-qubit gates per step, which kept noise low enough for the signal to survive.

String breaking in two dimensions

String breaking is what happens when confinement reaches its limit. Stretch the flux tube between two charges far enough and it snaps. Field energy converts into new particle-antiparticle pairs. In QCD, this is how mesons decay and jets hadronize after an LHC collision. Simulating string breaking classically in more than one spatial dimension is intractable: it involves a far-from-equilibrium process with exponentially growing entanglement.

In June 2025, two independent groups published string-breaking results side by side in Nature.

Superconducting Cochran and collaborators implemented a Z₂ gauge theory on a 2D lattice of Google superconducting qubits. They imaged the string connecting two charges as confinement strength varied. They also identified a resonance where string breaking is enhanced.

Neutral atoms On QuEra’s Aquila platform, González-Cuadra, Zoller, and collaborators arranged rubidium atoms in a kagome optical-tweezer array that naturally encodes a confining gauge theory. They watched the flux tube form between charges and snap to create new particle pairs: the first observation of string breaking in a 2D quantum simulator.

Two-dimensional lattice QED, with matter, on a qudit processor

Lattice gauge simulations face a representation problem. A gauge field at each link on the lattice has an infinite-dimensional spectrum. Encoding that in two-level qubits requires severe truncation and many qubits per link. Qudits offer a more natural path: a d-level quantum system can represent d field states in a single particle.

Trapped ions In April 2025, Meth and colleagues at Innsbruck published the first quantum simulation of 2D lattice QED with both dynamical gauge fields and matter, in Nature Physics. They used a trapped-ion qudit processor. Gauge fields were encoded in up to seven Zeeman levels of calcium-40 ions; matter sat in standard qubits. The simulation computed ground-state properties of a QED plaquette. Increasing the qudit dimension captured physics invisible at the qubit level, and the approach cut circuit depth compared with equivalent qubit implementations.

Hadron scattering on quantum hardware

Scattering experiments are the bedrock of particle physics. Almost everything we know about the subatomic world comes from colliding particles and reading the debris. Simulating scattering on a quantum computer means preparing wave packets, evolving them through a gauge field, and extracting the resulting particle states. In May 2025, two groups independently did exactly this.

Superconducting Schuhmacher and collaborators at IBM Research and the University of Freiburg observed hadron scattering in a Z₂ lattice gauge theory on a digital quantum computer. They created meson-like bound states, collided them, and tracked the post-collision dynamics.

Trapped ions Independently, Davoudi, Hsieh, and Kadam at Maryland prepared meson wave packets on IonQ’s Forte hardware using up to 27 qubits and simulated their collisions.

Both results are preprints. Neither approaches real QCD complexity. But they establish the full workflow: state preparation, time evolution through a gauge theory, and observable extraction.

Quantum sensors are already searching for new particles

The intersection of quantum technology and particle physics is not limited to computation. Quantum sensing has begun to reshape the search for dark matter, and the results are no longer projections. The same precision-measurement logic now drives the hunt for neutrino mass.

In April 2025, the QUAX collaboration published in Physical Review X the first axion dark matter search using a superconducting transmon qubit as a single-photon counter. Placed inside a haloscope cavity, the qubit bypasses the standard quantum limit on conventional amplifiers. The result: a 20-fold improvement in scan speed. In January 2026, a Chinese collaboration reported in Nature the first intercity quantum sensor network for dark matter. Five nuclear-spin sensors were distributed across Hefei and Hangzhou, 320 kilometres apart, searching for transient spin rotations that axion-like particles would induce. The network set new constraints on ultralight axion couplings in a previously unexplored mass range.

The HAYSTAC experiment at Yale reported the broadest axion search to date using quantum squeezing to suppress noise. A December 2025 Physical Review Letters paper proposed using phase differences between distributed sensors to track the dark matter wind’s direction. One theoretical result tempers the enthusiasm. In April 2026, Rodd and collaborators at Berkeley showed in Physical Review Letters that intrinsically quantum effects of axion dark matter are washed out by mode averaging and noise. A classical-field treatment remains reliable for realistic detectors. The quantum advantage here comes from the sensors, not from the dark matter itself being quantum.

How far from real QCD

Every result above uses a simplified gauge theory. The gap between Z₂ or U(1) models and full SU(3) QCD is enormous. A Z₂ theory has two field states per link. QCD has a continuous, non-abelian gauge group with eight gluon fields, three colour charges, and fermionic quarks. The most advanced resource estimates call for thousands of logical qubits and trillions of gate operations for even modest lattice volumes. A useful QCD simulation sits well beyond any current or near-term hardware.

What has changed is the trajectory. Two years ago, quantum gauge-theory simulations were confined to one spatial dimension. Today, three platforms (superconducting qubits, trapped ions, neutral atoms) have demonstrated gauge-theory dynamics in 2D. Confinement, string breaking, and scattering have all been observed on hardware. The path from Z₂ to SU(2) to SU(3) is a gradient, not a cliff, and the field is climbing it. Davoudi’s review, presented at Lattice 2025 and posted as a preprint in May 2026, catalogues the full programme.

References

• J. Mildenberger et al., Confinement in a Z₂ lattice gauge theory on a quantum computer, Nature Physics 21, 312–317 (2025). Preprint: arXiv:2203.08905
• T. A. Cochran et al. (Google Quantum AI), Visualizing dynamics of charges and strings in (2+1)D lattice gauge theories, Nature 642, 315–320 (2025). Preprint: arXiv:2409.17142
• D. González-Cuadra et al. (Innsbruck, Harvard, QuEra), Observation of string breaking on a (2+1)D Rydberg quantum simulator, Nature 642, 321–326 (2025)
• M. Meth et al., Simulating two-dimensional lattice gauge theories on a qudit quantum computer, Nature Physics 21, 570–576 (2025). Preprint: arXiv:2310.12110
• J. Schuhmacher et al., Observation of hadron scattering in a lattice gauge theory on a quantum computer, preprint arXiv:2505.20387 (May 2025). Not peer-reviewed.
• Z. Davoudi, C.-C. Hsieh, S. V. Kadam, Quantum computation of hadron scattering in a lattice gauge theory, preprint arXiv:2505.20408 (May 2025). Not peer-reviewed.
• QUAX Collaboration, Quantum-enhanced sensing of axion dark matter with a transmon-based single microwave photon counter, Physical Review X 15, 021031 (2025)
• X. Bai et al. (HAYSTAC Collaboration), Dark matter axion search with HAYSTAC Phase II, Physical Review Letters 134, 151006 (2025). Preprint: arXiv:2409.08998
• H. Fukuda et al., Directional searching for light dark matter with quantum sensors, Physical Review Letters 135, 241802 (2025)
• Y. Wang et al., Constraints on axion dark matter by distributed intercity quantum sensors, Nature 650, 314–319 (2026)
• N. Rodd et al., Intrinsically quantum effects of axion dark matter are undetectable, Physical Review Letters 136, 171601 (2026)
• Z. Davoudi, Quantum simulation of gauge theories for particle and nuclear physics (review), preprint arXiv:2605.20417 (May 2026). Not peer-reviewed.