Our digital world runs on bits: zeros and ones carried by electrical signals or pulses of light. Quantum technologies replace them with qubits, which can exploit entanglement, the phenomenon Einstein famously called “spooky action at a distance”. In today’s quantum technologies, a qubit is usually stored either in an internal property of a particle, such as the polarisation of a photon or the spin of an electron, or in bosons, the family of particles to which photons belong. But nature offers another family of particles: fermions, of which the electron is the best-known example. Fermions obey peculiar rules. Two identical fermions can never occupy the same state, and swapping two of them flips the sign of their joint description. The QINF project explores what happens when information is encoded not in a property of a particle, but simply in the presence or absence of a fermion: a new unit of information we call the “febit”. The question is easy to state. In a network where each node can only talk to its neighbours, like computers linked by cables or atoms held in a row of laser traps, can febits do things that no number of qubits can? A first result from our team shows that, in a simple loop network, they can.

Going further would complete a three-step ladder. Since Bell’s theorem in 1964, we have known that qubits can produce correlations that no number of bits can reproduce. QINF aims to show that febits, in turn, can produce correlations that no number of qubits can reproduce, so that each step up, from bits to qubits to febits, is strictly more powerful than the one before. To get there, the project will develop the mathematics and computer algorithms needed to determine what qubits and febits can and cannot achieve in large networks. It will also tackle a very concrete problem. In large networks, the main bottleneck is often not computing power but the time spent waiting for messages to travel between nodes. QINF will identify basic network tasks, such as sharing resources between users or routing messages, where exchanging qubits instead of bits reduces the number of communication rounds needed, and prove where no such gain is possible. These results will help design the quantum internet of tomorrow.

Left: in Bell’s scenario, a central source reaches every party. It can be used to demonstrate that qubits outperform bits.
Right: in a loop network, there is no central source and each node only communicates with its neighbours. QINF aims to use such networks (the loop network and more general networks) to demonstrate that fermions can produce correlations that qubits cannot reproduce, assuming limits on communication time. QINF builds the theory of such advantages.

Publications

Fermions are fundamentally more nonlocal than bosons
Fatemeh Moradi Kalarde, Sadra Boreiri, Xiangling Xu, Lucas Tendick, Salman Beigi, Paolo Perinotti, Tommaso Guaita, Marc-Olivier Renou
arXiv:2606.12363, 2026
Indistinguishable fermions sent through a quantum network can produce correlations that distinguishable particles or bosons cannot reproduce without extra communication. As a consequence, fermions strictly beat every qubit-based protocol for some distributed computing tasks.

In the news

Team and positions

QINF is carried out within the PhIQuS team, joint between Inria Saclay and École polytechnique, and will fund two PhD students and three postdoctoral researchers. Openings are announced here and on Quantiki. To discuss joining, write to marc-olivier.renou@inria.fr.

Funding

This work is supported by ERC European Union grant QINF.