Researchers have experimentally demonstrated that exotic quantum objects known as non-Abelian anyons can perform the full set of operations needed for universal quantum computing.
Using 54 qubits on Quantinuumโs H2 trapped-ion processor, scientists created these unusual collective quantum states and manipulated them through two operations known as braiding and fusion.
The combination allowed the researchers to demonstrate a universal set of quantum operations, meaning, in principle, the system can perform any quantum computation.
The result is particularly significant because non-Abelian anyons may also offer a route toward fault-tolerant quantum computing that avoids some of the most resource-intensive methods currently proposed for quantum error correction.
The research involved scientists from the University of Chicago Pritzker School of Molecular Engineering, Harvard University, Stony Brook University and Quantinuum and was published in Nature.
What Universal Quantum Computing Means
A useful general-purpose quantum computer must eventually be capable of running many different kinds of quantum algorithms.
Today's experimental quantum processors can perform sophisticated operations, but constructing machines that are both universal and resistant to errors remains one of the biggest challenges in the field.
A universal quantum computer requires a collection of operations known as a universal gate set.
Just as a small collection of basic logical operations can be combined to perform essentially any conventional computation, a universal quantum gate set allows researchers to construct any desired quantum operation.
The new experiment demonstrates such a set using non-Abelian anyons.
What Are Non-Abelian Anyons?
Despite the name, the anyons used in the experiment were not ordinary particles moving independently through space.
Instead, researchers created them as emergent quantum objects inside a quantum processor.
Many conventional qubits were entangled together so that their collective behaviour followed the mathematical rules associated with non-Abelian anyons.
This means a group of ordinary physical qubits can effectively behave like an entirely different type of particle.
These emergent objects possess unusual properties that make them attractive for quantum information processing.
Why They Are Different From Ordinary Qubits
Conventional qubits encode information using two quantum states โ commonly represented as 0 and 1 โ along with superpositions of those states.
Non-Abelian anyons store quantum information differently.
Their information is distributed across the collective quantum state of multiple particles rather than being confined to one local object.
This nonlocal encoding can make information naturally less sensitive to certain small disturbances.
That property has long made anyons attractive candidates for topological quantum computing, an approach intended to protect quantum information through the structure of the quantum state itself.
Braiding Anyons to Perform Computation
One of the most unusual properties of non-Abelian anyons involves an operation known as braiding.
Imagine moving one anyon around another.
In ordinary physics, swapping identical particles may make little difference.
For non-Abelian anyons, however, the sequence in which particles move around one another matters.
Different braiding sequences transform their shared quantum state in different ways.
Those transformations can act as quantum computational gates.
In other words, computation can be performed by controlling how these emergent objects move around one another.
The Order of Operations Matters
The term โnon-Abelianโ refers to the mathematical property that changing the order of operations can change the result.
If operation A is performed before operation B, the final quantum state may differ from what would happen if B occurred before A.
This provides a rich set of possible transformations for quantum information.
Researchers have therefore spent years investigating whether these particles could form the basis of powerful and naturally protected quantum computers.
But an important limitation remained.
Braiding Alone Was Not Universal
In previous work published in 2024, researchers including University of Chicago scientist Ruben Verresen created non-Abelian anyons on Quantinuum quantum hardware.
Those anyons were associated with a mathematical symmetry group known as D4, describing the rotations and reflections that leave a square unchanged.
The experiment showed that non-Abelian order could be created and manipulated on a programmable quantum computer.
However, braiding those anyons could not generate every operation required for universal quantum computation.
The system was therefore computationally interesting but incomplete.
Switching From a Square to a Triangle Symmetry
For the new experiment, the researchers turned to another mathematical structure known as S3 symmetry.
S3 describes transformations associated with the permutations of three objects and can be related geometrically to the rotations and reflections of an equilateral triangle.
The corresponding anyons possess computational properties different from those associated with the earlier D4 system.
Researchers created these S3 non-Abelian anyons using 54 entangled qubits on Quantinuum's H2 processor.
But one crucial ingredient was still needed.
Fusion Provides the Missing Operation
The researchers combined braiding with another process called fusion.
Fusion occurs when two anyons are brought together and their combined quantum state is measured.
Different pairs of anyons can fuse in different ways, producing different measurable outcomes.
Those outcomes contain quantum information and can be used as computational operations.
By combining fusion measurements with braiding, the researchers gained access to operations unavailable through braiding alone.
This proved to be the key to achieving universality.
Three Operations Were Enough
The researchers demonstrated three essential tools.
One was an entangling gate generated through braiding.
The other two were different types of measurements performed through fusion.
Together, these operations form a set capable, in principle, of generating arbitrary quantum computations.
This is what allowed the researchers to describe the system as having a universal gate set.
The experiment therefore moved non-Abelian anyons from demonstrating interesting topological operations to demonstrating the core ingredients necessary for general-purpose quantum computation.
Quantum Information Stored in Qutrits
The researchers encoded information using pairs of anyons to create topological qutrits.
A conventional qubit has two basic states.
A qutrit has three.
This additional level changes the mathematical structure of the computation and can provide access to operations unavailable in simple two-level systems.
The topological qutrits in the experiment were encoded collectively across the anyonic system rather than within single physical components.
An Idea Proposed More Than 20 Years Ago
The theoretical foundation for this strategy dates back to work proposed in 2003 by Carlos Mochon, then a student of physicist John Preskill at Caltech.
Mochon's work showed mathematically how suitable non-Abelian anyons, combined with operations beyond straightforward braiding, could support universal quantum computation.
Turning that abstract theoretical idea into an experiment required more than two decades of advances.
Quantum hardware needed to improve enough to create, manipulate and measure sufficiently complicated entangled states.
The new experiment demonstrates that those ideas can now be implemented on actual programmable quantum hardware.
The Error-Correction Problem
Universal computation is only one challenge facing quantum computers.
Another is error correction.
Quantum information is extraordinarily fragile.
Interactions with the environment, imperfect controls and measurement errors can destroy the delicate quantum states needed for computation.
One common strategy is to encode one logical qubit across many physical qubits.
The redundant structure allows errors to be detected and corrected.
But protecting the quantum information creates another problem.
Some computational operations are difficult to perform directly on error-corrected logical qubits.
Why Magic States Are So Expensive
To perform operations missing from many standard error-correcting codes, quantum computing architectures often rely on special resources known as magic states.
These states allow otherwise inaccessible quantum gates to be implemented.
Unfortunately, creating high-quality magic states typically requires a process called magic-state distillation.
Many imperfect states are consumed to produce a smaller number of highly reliable ones.
This procedure can require enormous numbers of physical qubits and could represent one of the largest resource costs in future fault-tolerant quantum computers.
Non-Abelian Anyons Could Offer Another Route
The new study suggests non-Abelian codes might reduce or avoid this requirement.
Because their computational structure is richer than that of many conventional quantum error-correction codes, they may support a wider collection of fault-tolerant operations directly.
Researchers demonstrated that their anyonic system could produce a magic state through topological operations.
This is significant because the state was generated without carrying out the conventional resource-intensive distillation procedure.
In principle, that could lead to substantially more efficient architectures.
Why Researchers Call It a โDark Horseโ
Quantinuum researcher Henrik Dreyer described non-Abelian codes as a โdark horseโ in the race toward quantum error correction.
Much of the quantum-computing community currently focuses on better-established approaches such as surface codes.
Non-Abelian codes remain less developed experimentally.
But they may offer computational capabilities that reduce some of the enormous overhead associated with conventional fault-tolerant architectures.
The new demonstration strengthens the case that this alternative deserves serious consideration.
But the Experiment Was Not Yet Fault Tolerant
An important limitation remains.
The researchers did not demonstrate active error correction in the experiment.
The goal was instead to establish whether the required computational building blocks could be implemented successfully.
The system should therefore be viewed as a proof of principle rather than a functioning fault-tolerant quantum computer.
The next major challenge is to combine these universal operations with active mechanisms that detect and correct errors while the computation is taking place.
Building Stable Non-Abelian Quantum Memories
Researchers are already working toward that next step.
Verresen and colleagues are exploring methods for stabilizing non-Abelian quantum memories.
A quantum memory must preserve quantum information reliably for long enough to perform useful calculations.
If non-Abelian states can be stabilized while also supporting universal operations, they could provide both protection and computation within the same topological framework.
That combination would be highly attractive for future large-scale quantum machines.
Why Topological Protection Could Matter
Conventional qubits can be extremely sensitive to local disturbances.
Because information in topological systems is spread across many components, a small error affecting one location may not immediately destroy the encoded information.
The important information is stored in the global relationship among the particles.
This is why topological approaches have long been considered promising for fault-tolerant quantum computing.
Non-Abelian anyons add another advantage: their braiding and fusion rules can themselves perform computations.
Creating Artificial Quantum Worlds
The research also illustrates a remarkable capability of modern quantum processors.
Scientists are not merely using quantum computers to calculate solutions to equations.
They are programming qubits so that their collective behaviour mimics exotic forms of matter.
Verresen compared such systems to creating small artificial universes with their own particle-like objects and interaction rules.
Within these engineered quantum environments, researchers can study forms of physics that are difficult to access in ordinary materials.
More Than a Computing Experiment
Non-Abelian anyons are therefore interesting beyond their potential technological applications.
They provide experimental platforms for investigating fundamental concepts in topology, quantum information and many-body physics.
By manipulating these states, researchers can test predictions that previously existed mainly as mathematical theory.
The line between quantum simulation and quantum computing becomes increasingly blurred: the same hardware can explore exotic physics while also demonstrating new computational architectures.
A New Candidate for Universal Quantum Computing
The experiment does not establish that non-Abelian anyons will become the dominant architecture for future quantum computers.
Major engineering challenges remain, especially active error correction, scalability and maintaining complex entangled states over long computations.
But the study removes an important theoretical and experimental obstacle.
Non-Abelian anyons are no longer merely a promising way to store protected quantum information.
Researchers have now experimentally demonstrated that they can support the full range of operations required for universal quantum computation when braiding and fusion are combined.
A Different Route Toward Fault-Tolerant Machines
Most approaches to quantum computing attempt to build reliability by adding increasingly sophisticated layers of error correction around fragile physical qubits.
The anyonic approach starts from a different idea.
Instead of protecting information only through redundancy, it attempts to encode that information in collective quantum structures whose topology already provides some protection.
If such systems can eventually perform universal operations while maintaining active error correction, they could reduce some of the enormous resource requirements expected for conventional quantum architectures.
For now, the demonstration marks an important proof of principle: an exotic form of quantum matter once considered mainly a theoretical possibility has performed the fundamental operations needed to become a serious candidate for universal quantum computing.
Journal reference
Chiu Fan Bowen Lo, Anasuya Lyons, Dan Gresh, Michael Mills, Peter E. Siegfried, Maxwell D. Urmey, Nathanan Tantivasadakarn, Henrik Dreyer, Ashvin Vishwanath, Ruben Verresen and Mohsin Iqbal. โUniversal gates from braiding and fusing anyons on quantum hardware.โ Nature, 2026, 655(8123), 591.