Braiding, Then Fusing: A 54-Qubit Reach for Universal Topological Quantum Computing
Quantum computers have a fragility problem. The quantum states that give them their power are easily disturbed, and a stray bit of noise can scramble a calculation before it finishes. One of the most elegant proposed cures is to hide the information where noise cannot easily reach it: not in any single particle, but in the pattern of how exotic particles called anyons are woven around one another. Braid two anyons, one past the other, and you change the system's state in a way that depends only on the topology of the braid, on which strand went over which. Small local errors leave that pattern intact. This is the promise of topological quantum computing: a qubit that is robust because its information is smeared across a shape rather than pinned to a point.
There has always been a catch. For the simplest, most experimentally tractable kinds of non-Abelian anyons, braiding alone is not enough to run every possible algorithm. The set of operations you can build by braiding is powerful but incomplete, short of what computer scientists call universality. To reach the full range, you need something more.
A team spanning Quantinuum and academic collaborators has now shown, in a paper published in Nature, what that something can be. The move is fusion. When two anyons are brought together they can merge, and the outcome of that merger is not always fixed; measuring which anyon results yields information and, with it, a new computational operation. Treated as a primitive in its own right, and combined with braiding, fusion closes the gap. The authors demonstrate that a minimally non-Abelian system can be made universal by using anyon fusion as a computational tool.
They did it not on paper but on hardware. The experiment ran on Quantinuum's H2 processor, using 54 qubits to prepare a topological state built on S3, the smallest non-Abelian group, whose anyons obey cyclic fusion rules. Within that state the team trapped a single non-Abelian anyon on a virtual torus, encoded logical information in the fusion space of the quasiparticles, and combined braiding operations with charge measurements to realize universal gates. They also prepared what is known as a magic state, a further ingredient of full-blown quantum computation, and read it out. The point of the exercise is that S3 topological order is at once simple enough to build today and rich enough, once fusion is added, to compute anything.
It is worth noting that this is not a discovery from the past few days. The work first appeared as a preprint arXiv:2601.20956 , submitted in Jan. 28, 2026, and revised in the spring. In July, the peer-reviewed version was published in Nature. This is an important milestone, as independent expert review matters for a bold experimental claim.
Topological quantum computing has long been the field's most beautiful idea and one of its most stubborn to realize, precisely because the anyons it depends on are hard to make and manipulate. Showing, on a working 54-qubit machine, that braiding plus fusion can in principle do everything a quantum computer needs to do is a concrete step from theory toward a usable device. It does not deliver a fault-tolerant topological computer, and the scale here is still that of a demonstration rather than a practical machine. But it narrows the distance between the elegant sketch in the textbooks and an instrument someone could one day compute on.