Q Neko, a new EU–Japan hybrid quantum classical project under the EuroHPC Joint Undertaking (EuroHPC JU), links Japan's ABCI Q national quantum testbed to leading EU supercomputers.
Europe and Japan have €4 million and 36 months to find out whether their different quantum hardware ecosystems can run the same hybrid quantum-classical software.
The Q-Neko project, which started January 1 under the EU-Japan Digital Partnership, will run that test publicly for the next three years. The European side contributes roughly €4 million from Horizon Europe, with project coordination handled by CSC, Finland's IT Center for Science. Japan co-funds through its Cross-ministerial Strategic Innovation Promotion Program and provides access to ABCI-Q, one of Japan's main quantum testbeds.
The core test is software alignment. Hybrid quantum-classical systems split work between a quantum processor and a conventional supercomputer, with the classical side handling orchestration, error correction, and pre- or post-processing. That division of labor only works if both sides agree on the same data formats, compiler targets, and benchmark definitions. EuroHPC JU's existing HPC+QC work has built pieces of this stack inside Europe; Q-Neko is the joint program that asks Japan to run on the same rails.
The consortium picked the application domains on purpose. Carbon capture, telecommunications, fluid dynamics, satellite image analysis, materials science, and quantum-enhanced machine learning each stress a different part of the hybrid stack: a CO₂ chemistry workload pushes the classical side differently than fluid dynamics, and a satellite imaging benchmark exercises different data movement. If the stacks are aligned, the same software should run on a leading EuroHPC system and on Japan's ABCI-Q without bespoke porting.
The history of cross-region quantum projects does not make this easier. Hardware vendors on both sides are still arguing about quantum instruction sets, error-correction schemes, and low-level compilers. There is no equivalent of the Linpack benchmark for hybrid systems, so the project has to draft its own definitions before any of the joint runs can be compared. €4 million is enough to fund coordination, expert exchanges, and the benchmark work. It is not enough to build new hardware on either side.
Q-Neko is also testing the collaboration model. The project funds researcher and engineer exchanges between the two regions, on the assumption that people who have worked with each other's code will spot integration problems faster than a remote working group. The 2026 and 2027 milestones on the project calendar, both of which involve joint benchmark runs, will be the first public signal of whether the exchange produces compatibility or just contact.
By 2028, the project will be judged on a concrete bar. The minimum is a published set of hybrid benchmarks that both consortia run without modification, plus at least one reference application whose results are reproducible on the other region's hardware. The maximum is a reusable hybrid software stack that survives contact with the next generation of quantum machines on both sides. Below the minimum, the €4 million will still fund three years of useful research. Above the maximum, the program becomes infrastructure.
Two watch items will show whether the program is on track. A first joint benchmark draft in late 2026 will show whether the two regions can agree on a common interface at all. A shared reference application reproducing earlier results by mid-2027 will show whether the integration is more than a working group. The carbon-capture and satellite-imaging use cases will read as either proof points or footnotes depending on which of those two milestones lands first.