Topological qubits
the high-risk bet on a qubit that protects itself
Key facts
- Error-resistantprotected by physics
- The idea
- Majoranasquasiparticles, contested
- Built on
- Microsofttwo decades of work
- Main backer
- None validatedMajorana 1 claim contested
- Working machine
Every other approach fights errors after they happen. This one tries to build a qubit that cannot easily be disturbed in the first place, by hiding the information in a knot. Decades of work, billions spent, and the physics is still argued over.
Every quantum computing platform fights the same enemy: the environment nudges a qubit and the information degrades. Most approaches accept that and spend enormous resources on error correction, using many physical qubits to protect one reliable logical qubit. The topological approach asks a different question. What if the qubit were built so the disturbance did not reach the information in the first place?
The idea
The information in a topological qubit would be stored not in the state of one particle but in a global property of a system, a property that only changes if you change the system’s overall arrangement. The usual analogy is a knot in a rope: prod the rope anywhere and the knot survives, because a knot is a property of the whole thing rather than any point on it.
The proposed physical basis is the Majorana zero mode, a quasiparticle predicted to appear at the ends of certain superconducting nanowires. A quasiparticle is not a fundamental particle but a collective behaviour of many electrons that acts like one. Information would be shared between two of these modes at opposite ends of a wire, so no local disturbance could read or corrupt it, and computation would proceed by braiding them around one another.
Why it would be worth it
The payoff is the reason serious money has gone into it for two decades. If the error protection is built into the physics, the overhead of error correction collapses. Where a superconducting machine may need a thousand or more physical qubits to make one dependable logical qubit, a topological machine might need a handful. A useful quantum computer could then be built at a fraction of the scale, which reframes the entire engineering problem.
Why it remains contested
The obstacle is that Majorana zero modes are extraordinarily hard to prove you have made. The experimental signature that indicates their presence can also be produced by mundane disorder in the material, and the field has a difficult history here: high-profile results claiming detection have been challenged, re-examined and in one prominent case retracted. Independent replication has been slow, and researchers disagree in good faith about what the data show.
Microsoft has been the most committed backer and continues to publish. In February 2025 it announced Majorana 1, a chip it described as the first quantum processor powered by a topological core, with eight topological qubits on a design intended to house a million. The announcement drew the same scrutiny as the field’s earlier claims: several physicists noted that the accompanying paper did not itself demonstrate a working topological qubit, and independent validation has not followed. Much of the rest of the field has meanwhile moved on to platforms that produce working, if imperfect, machines today.
Where it stands
Topological quantum computing remains the highest-risk, highest-reward approach in the field. There is no topological quantum computer running useful circuits, and no undisputed demonstration of the braiding operations the design depends on. It is the one platform where the question is still whether the physics works at all, rather than how quickly the engineering can be scaled.