TL;DR
Alice & Bob reported that it tested core components of its cat-qubit error-correction architecture on three hardware demonstrators for Phase 1 of France’s PROQCIMA program. The tests included protected CNOT gates, repeated error-correction cycles and a non-Clifford gate, but the measured logical error rate has not fallen below the error-correction threshold.
French quantum-computing company Alice & Bob says it demonstrated core operations of its cat-qubit error-correction architecture on three hardware demonstrators during Phase 1 of the national PROQCIMA program. The tests included protected CNOT gates, repeated error-correction cycles and a non-Clifford gate, but the company reported that its measured logical error rate remains above the error-correction threshold needed for fault-tolerant operation.
In a company summary released October 20, Alice & Bob said it assembled and operated two six-qubit chips and one 12-qubit chip. The demonstrators were built under a Phase 1 commitment to integrate and partially validate the main operations in its error-correction architecture on real hardware. The company said it documented measured performance for each building block, with further optimisation under way.
Reported tests included noise-biased CNOT gates between cat qubits, designed to preserve their built-in protection against bit-flip errors. Alice & Bob also reported a measure-and-reset sequence for auxiliary qubits, which reads out error-syndrome information and prepares the qubits for another correction cycle, and multiple cycles of a distance-three repetition code using three data qubits. The company said performance of the measure-and-reset sequence was broadly in line with program targets.
Alice & Bob also reported measuring logical error across the full register and demonstrating a non-Clifford gate, a component needed for universal fault-tolerant quantum computing. Those demonstrations do not mean the system has reached fault tolerance: the company’s own account says the logical error rate is still above threshold. The announcement did not provide a numerical logical error rate or a detailed comparison with the threshold.
The Gap Between Demonstration and Fault Tolerance
The results matter because they show the company has tested several connected components of its proposed architecture on physical processors, rather than describing a design solely through simulations. For quantum computers to correct errors effectively, operations must work together across repeated cycles, and the resulting logical error rate must fall below a relevant error-correction threshold. Alice & Bob says it has demonstrated some of those operations, while also reporting that the threshold has not yet been reached.
The gap is central to the program’s longer-term goal: PROQCIMA aims to develop two universal quantum-computer prototypes, each capable of supporting 1,000 logical qubits, for industrialisation by 2032. Phase 1 results are an interim technical milestone, not evidence that either prototype is ready. The company’s report points to coherence consistency across the chip as a priority for improving performance, but does not quantify how much improvement would be needed or when it might be achieved.
The announcement also highlights the engineering work required to keep quantum hardware stable. Alice & Bob says automated monitoring and recalibration software tracked more than 50 indicators per qubit and ran while the quantum processing unit was idle. The company says the software can identify drift and support recalibration; it reported that manually detecting and correcting drift could previously cost up to a day of operating time. These are company-reported operational details, not an independent assessment of system performance.
quantum computing error correction kit
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
France’s Ten-Year PROQCIMA Competition
PROQCIMA is a French government quantum-computing program led by the French Defense Innovation Agency (AID) and the Direction générale de l’Armement. According to Alice & Bob’s announcement, it is structured as a 10-year competition intended to advance candidate approaches toward two universal quantum-computer prototypes. Five companies were selected to take part: Alice & Bob, Pasqal, Quandela, Quobly and C12.
The program’s first phase was designed to evaluate the technical and industrial feasibility of each candidate’s approach, including maturity demonstrated on real hardware. Alice & Bob said it began its Phase 1 work in 2024 and committed to an integrated multi-block demonstrator with partial functional validation. Under the program structure described by the company, three candidates are to continue after the first four years, and two are ultimately to be selected to continue beyond year eight. The announcement does not identify the outcome of those later selections.
Alice & Bob’s approach uses cat qubits, which are intended to make one type of error less likely than another through their physical encoding. The company’s reported noise-biased gates and repetition-code tests relate to that strategy. The results summarized here are the company’s account of its own work; the supplied material does not include an independent technical review of the measurements.
“The race for fault-tolerant quantum computing is a question of who will master one of the defining technologies of this century, and France starts from a position of real strength.”
— Théau Peronnin, CEO of Alice & Bob
As an affiliate, we earn on qualifying purchases.
Threshold Performance Still Unreported
The company says the measured logical error rate is not yet below threshold, but its announcement does not state the numerical rate, the threshold value used, or the measurement conditions needed to compare the two. It also does not give detailed results for each demonstrator, including how performance varied between the six-qubit and 12-qubit chips.
Alice & Bob’s digital-twin simulations identified more consistent, longer qubit coherence across the chip as a priority for operating below threshold. The supplied account does not specify a target coherence time, the expected schedule for reaching that target, or whether other hardware or control improvements will be required. Independent validation of the reported measurements is not included in the material provided.
The program’s eventual selection decisions and the timing of later milestones are also not detailed in this results summary. The Phase 1 announcement describes work completed by Alice & Bob, but does not report how the results compare with those of the other participating companies or how program officials assessed them.
quantum error correction demonstration kit
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Coherence Work and Program Reviews
Alice & Bob says optimisation of the cat-qubit building blocks is continuing, with more consistent qubit coherence across the chip identified as an early priority by its digital-twin simulations. The company has not announced a date for reaching below-threshold logical error rates or provided a schedule for the next hardware results in the supplied material.
PROQCIMA’s next broader milestones will include decisions about which candidates continue after the program’s first four years, followed by a later selection of two candidates beyond year eight, according to the structure described by Alice & Bob. The program’s stated destination is two 1,000-logical-qubit prototypes ready for industrialisation by 2032. How the Phase 1 assessments will affect candidate selection, and what benchmarks will govern those decisions, remains unreported in this announcement.
fault-tolerant quantum computing components
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What did Alice & Bob report demonstrating?
The company said it tested core operations of its cat-qubit error-correction architecture on three demonstrators: two six-qubit chips and one 12-qubit chip. Reported tests included protected CNOT gates, measure-and-reset operations, repeated distance-three repetition-code cycles and a non-Clifford gate.
Has Alice & Bob reached fault-tolerant quantum computing?
No. The company reported that its measured logical error rate remains above the error-correction threshold. It did not provide a numerical rate or threshold value in the supplied results summary.
What is PROQCIMA aiming to produce?
The French government program aims to develop two universal quantum-computer prototypes, each able to support 1,000 logical qubits, for industrialisation by 2032, according to Alice & Bob’s account.
What does Alice & Bob say must improve?
The company said simulations using its digital twin point to the need for longer, more consistent qubit coherence across the chip to operate below the error-correction threshold. It has not specified a target or completion date.
Who else is participating in PROQCIMA?
Alice & Bob named Pasqal, Quandela, Quobly and C12 as the other companies selected for the program. Its announcement says three candidates are to continue after the first four years, with two selected to continue beyond year eight.
Source: rss
