Non-Abelian Anyons Can Run Any Quantum Computation, Chicago-Led Team Says

The experiment, published in Nature and announced by the University of Chicago on 25 September, combined braiding and fusion to produce a universal gate set. ScienceDaily, which carried the university's release, noted the work involved UChicago's Pritzker School of Molecular Engineering, Harvard, Stony Brook University and Quantinuum.

Universal means a machine can run any quantum algorithm, not just one specialised task. The result is a proof of principle on 54 qubits, not a working general-purpose computer.

Error correction is the field's central engineering problem. Quantum computers are fragile: information is typically spread across many physical qubits to survive noise, but those protected states usually cannot execute every operation a universal machine needs. Engineers fill the gap with "magic states", prepared resources that demand a purification step called distillation. This process can eat a large share of a machine's available qubits. The new paper argues non-Abelian anyons may avoid that cost entirely.

"Non-Abelian codes are a dark horse in the race to quantum error correction," said Henrik Dreyer, managing director and scientific lead at Quantinuum's Munich office and a co-author of the study. "In this work we show the first universal gate set in a non-Abelian code, which demonstrates that fault-tolerant computations can in principle be done without resorting to magic state distillation or cultivation, which are the most expensive operations in standard quantum error correction codes."

Anyons do not exist as standalone particles in nature. Scientists create them inside circuits by entangling many ordinary qubits into a collective state that obeys its own rules. Moving one anyon around another, a process called braiding, changes the encoded state. The order of those moves matters, which is what "non-Abelian" means. Because the information is spread across the entangled group, it is also shielded from some small disturbances.

An earlier 2024 experiment by some of the same researchers created anyons tied to a symmetry group called D4, the rotations and reflections that leave a square unchanged, on a Quantinuum trapped-ion machine. That showed the particles could be made and moved, but braiding alone was not enough for universal computation.

"In that work, we didn't demonstrate that those emergent forces were enough to do quantum computation," said Ruben Verresen, assistant professor of molecular engineering at UChicago PME and a co-author. "That particular universe we created was not powerful enough."

For the new study the team switched to S3, the symmetry group of an equilateral triangle, and built the corresponding anyons on Quantinuum's H2 processor using 54 entangled qubits. Verresen described the codes as "creating little universes", in comments released by the university. The paper's claim is narrow and testable: this particular code supports a universal gate set in hardware.

The same week, the US Department of Energy addressed a different question: what quantum computers should be used for. On 25 September Fermilab published the SCAC Quantum Committee Report, "Path to an Integrated Quantum Future", following a field-wide effort chaired by Fermilab chief technology officer Anna Grassellino, with University of Chicago professor Supratik Guha as vice chair.

The report sets a milestone-driven path toward a scientifically relevant, error-corrected quantum computer by 2028. It sketches a three-phase plan: competitive Grand Challenges from 2026 to 2028 pairing national labs, universities and industry; a DOE Quantum Computing User Facility, described in the report as an open scientific instrument rather than a commercial cloud service; and integration of quantum processors into DOE high-performance computing and AI infrastructure from 2030 onward. The stated target applications include molecular prediction for drug discovery, catalysts for manufacturing, fusion-relevant materials and early-universe physics.

Hardware makers are pushing on the cryogenic side in parallel. IBM said on 19 August that it had joined and cooled two modular cryogenic systems into a single environment, a step toward linking hundreds of quantum chips. The modules stand more than 8 feet tall and 8 feet wide, cool to 4 Kelvin in under five days and reach below 15 millikelvin. IBM describes this as more than 180 times colder than deep space. The company plans to install Nighthawk processors in the modules later this year and still targets a fault-tolerant machine, Starling, in 2029.

Between the qubits and the fridges sits a less visible problem: modelling the control and readout electronics that must work a few degrees above absolute zero. A sponsored article published by Semiconductor Engineering on 24 September describes a workflow that extracts a BSIM-BULK compact model at room temperature. It then uses a hybrid neural network to learn the residual cryogenic behaviour the physics-based model cannot capture, using measurement data at 4 K or below. The article cites a 41.8% compound annual growth rate for the quantum computing market from 2025 to 2030, sourced to MarketsandMarkets. This is a vendor-side projection, not a measurement.

The two September announcements point in opposite directions. The anyon result is a specific, hardware-verified claim about one route to fault tolerance. Its authors are explicit that it removes one expensive step in principle rather than in a production machine. The DOE roadmap is a policy document with dates and no new experiment behind it.

One disagreement stands out: the SCAC report argues success should be measured by scientific utility rather than hardware-centric milestones, while the same roadmap still commits to an error-corrected machine by 2028. Those two measures are not the same, and the report does not reconcile them.

The near-term test is whether other groups can reproduce the S3 universal gate set and whether fusion-based operations scale beyond 54 qubits. Until then, the dark horse has won a race, not the race.