Superconducting qubits have a well-known weakness: they are charged circuits, giving electromagnetic noise a built-in coupling channel. As qubit counts scale up, this channel is amplified with them—today one of the main obstacles to scalability.
In October 2026, the University of Surrey's quantum science group published a design paper in *npj Quantum Information* with a counterintuitively simple idea: remove the charge. The proposed device is called the Superfluid Helium Oscillator Quantum (SHOQ), based on superfluid helium-3—a neutral liquid that flows without friction. The authors are Priya Sharma and Eran Ginossar of the University of Surrey, together with Jens Koch of Northwestern University. Koch co-developed the transmon, the most widely used superconducting qubit design today.
Neutrality as protection
The key property of superfluid helium-3 is electrical neutrality. A superconducting circuit is charged, so electromagnetic noise couples in directly; a neutral fluid offers no such handle. If the qubit's core degrees of freedom carry no charge, the noise sources that dominate superconducting designs lose their grip.
The described hardware: a cylindrical container housing a Josephson weak link, a nanopore that lets superfluid flow between two helium reservoirs, and a flexible membrane that responds to pressure changes. Energy flows back and forth between the weak link and the membrane; the quantized oscillations form the qubit's discrete states. The operating point is around half a millikelvin—below the superfluid transition temperature of helium-3-B—at micron-scale dimensions.
A side benefit is temperature: the cryogenic conditions for superfluid helium-3 have been achieved in prior experiments, so the challenge lies not in cooling but in fabricating the nanoscale weak link and membrane, and in reading out the quantum state.
It must be stated immediately: this entire branch rests on calculations and modeling. No prototype exists.
Where the "100x" comes from
The headline figure is that error rates "could be about 100 times lower than conventional superconducting qubits." The provenance matters:
- The paper's abstract in *npj Quantum Information* does not itself print the 100x comparison. It gives the device's electrical neutrality, micron-scale size, millisecond-scale coherence times, and the millikelvin regime required for superfluidity.
- The 100x figure appears in the University of Surrey's official announcement, worded as a computational prediction that the error rate would be about one percent of conventional superconducting qubits.
- It is therefore a theoretical prediction, benchmarked against current error rates of that class of superconducting devices—not a measurement from a SHOQ chip.
- Peer-reviewed and published: the device's physical design, parameters, and theoretical modeling (*npj Quantum Information* is a formal journal, not a preprint).
- The device exists as a blueprint, not hardware. The prototype is at the planning stage.
- The ~100x error-rate reduction is a theoretical prediction, experimentally unverified.
- Millisecond coherence is a target parameter, likewise unmeasured.
- The university announcement is institutional self-reporting; the 100x figure traveled via the announcement rather than the paper's abstract—read with that distance in mind.
- The extreme cryogenic conditions required for superfluid helium-3 have been achieved before, lowering the cold-end difficulty of any experiment.
The millisecond coherence target is an independent metric stated in the paper—but since the device exists only on paper, fabrication defects and readout electronics have not yet tested these predictions.
The authors are candid. In Sharma's words: she was not the first to think of the individual components behind the idea, but the team is the first to integrate these elements into a microfluidic device and work out the specific details needed for it to operate as a qubit. The math says it should work; the next step is a prototype to test the predictions.
Not meant to replace superconducting qubits
Ginossar's framing carries no replacement claim: one does not necessarily need a single qubit type to do everything—combining different quantum technologies lets each advantage be exploited. A longer-term possibility is SHOQ as quantum memory: the superfluid qubit stores quantum information while other hardware performs computation.
Neither use depends on beating superconducting hardware on every dimension. The announcement's two directions are also integration with existing superconducting circuits, and quantum memory. This makes the priority claim of "the first qubit design scheme based on superfluids" read more like opening a new route than announcing a replacement.
Koch's involvement matters practically: since SHOQ's performance benchmark is the transmon itself, the transmon's co-developer signed onto the design paper—the comparison standard was taken seriously at least at the conceptual level.
Prototype funding is in place: an IAA commercialization fellowship awarded to Sharma supports prototype development.
Where the proposal stands today
A detail worth noting separately: each component of the idea had been considered before by others. As an engineering move this is unglamorous, yet uncommon in quantum computing.
Viewing "all the parts exist" together with "there is finally a blueprint a lab can be handed" says more about the nature of this work than the 100x figure does. It discovered no new material and set no new world record. It gathered scattered possibilities into a specification sheet and handed it to the fabrication shop.