What the partners are building
France's CEA and Alice & Bob announced a research collaboration on 17 September to connect the startup's quantum software stack with existing high-performance-computing environments. The first integration uses Bull Qaptiva, the hardware-independent quantum application platform developed by Eviden, so workloads can be orchestrated between classical supercomputers and future Alice & Bob quantum backends.
The immediate goal is software and workflow integration rather than a new production quantum computer. CEA says Qaptiva will provide the hybrid execution layer and that its own teams will contribute expertise in coupling quantum processors with HPC systems. Reuters reports that the partners want to improve this French software path as an alternative to relying exclusively on Nvidia's CUDA-Q ecosystem for hybrid quantum-classical workloads.
Research starts with many-body physics
In parallel, the partners are starting a joint research programme on many-body problems in condensed-matter physics. They plan to identify algorithms that could run realistically on early fault-tolerant quantum systems around 2030, with possible applications in materials discovery, chemistry and complex industrial simulation.
That timetable is a research target, not a deployment commitment. Alice & Bob is developing cat-qubit hardware intended to suppress one class of quantum error, but the collaboration does not demonstrate a fault-tolerant machine capable of those industrial workloads today. The work instead tries to prepare software and candidate use cases before such hardware reaches the required scale.
Why the HPC link matters
Useful quantum computing is likely to operate alongside classical systems rather than replacing them outright. In the model described by CEA and Alice & Bob, a conventional supercomputer handles the broader workload and sends only suitable subproblems to a quantum processor. That makes scheduling, data movement and a common software interface part of the practical bottleneck, alongside qubit quality and error correction.
CEA already hosts quantum processors alongside classical HPC infrastructure and France agreed in June to acquire an 18-cat-qubit Alice & Bob system for installation in 2027. The new collaboration therefore extends an existing national integration programme from hardware procurement toward the software layer that researchers would use to combine the machines.
What remains unproven
The announcement establishes the collaboration, integration target and research programme, but it contains no benchmark showing a quantum advantage, no production performance result for the planned Alice & Bob system, and no evidence that the targeted many-body algorithms will outperform classical methods at useful scale. Those questions depend on hardware and algorithms that are still under development.
The material next evidence will be a working Qaptiva integration, published algorithm results, details of the 2027 system after installation, and eventually fault-tolerant hardware capable of running the proposed workloads. Until then, the programme is best understood as infrastructure preparation for hybrid quantum-HPC computing rather than a claim that the underlying quantum-computing threshold has already been crossed.