The characterization of the system indicated a few main limits, although. Cat qubits make bit flips extraordinarily unlikely, however not not possible. By focusing error correction solely on section flips, any bit flips that do happen inescapably set off the failure of the complete logical, error-corrected qubit. “Reaching lengthy logical bit-flip instances is difficult as a result of any single cat qubit bit flip occasion in any a part of the repetition code straight causes a logical bit flip error,” the authors observe. The opposite challenge is that the transmons used for error correction nonetheless endure from each bit and section flips, which may additionally mess up the complete error-corrected qubit.

The place does this go away us?

There are a variety of firms like Amazon which might be betting that utilizing a by some means much less error-prone {hardware} qubit will permit them to get efficient error correction utilizing fewer complete {hardware} qubits. In the event that they’re appropriate, they will be capable to construct error-corrected quantum computer systems utilizing far fewer qubits, and so probably carry out helpful computation sooner. For them, this paper is a crucial validation of the concept. You are able to do a type of mixed-mode error correction, with a strong {hardware} qubit paired with a compact error-correction code.

However past that, the messages are fairly combined. The {hardware} nonetheless needed to depend on much less sturdy {hardware} qubits (the transmons) to do error correction, and the very low error fee was nonetheless not low sufficient to keep away from having occasional bit flips. And, in the end, the error fee enhancements gained by rising the dimensions of the logical qubit aren’t on a trajectory that might get you a helpful degree of error correction with no need an unrealistically giant variety of {hardware} qubits.

In brief, the underlying {hardware} is not presently adequate to allow any type of advanced calculation, and it could want radical enhancements earlier than it may be. And there is not an apparent alternate path to efficient error correction. The potential of this method continues to be there, nevertheless it’s not apparent how we will construct {hardware} that lives as much as that potential.

As for Amazon, the image is even much less clear, on condition that that is the second qubit expertise that it has talked about publicly. It is unclear whether or not the corporate goes to go all-in on this method, or continues to be on the lookout for a expertise that it is prepared to decide to.

Nature, 2025. DOI: 10.1038/s41586-025-08642-7  (About DOIs).