A New Architecture for Photonic Quantum Computing
Researchers have introduced a modular, extensible architecture for photonic quantum computing called Clavina. This design solves a long-standing constraint in the field, where photonic processors were largely limited to linear operations. By integrating a central control unit with plug-and-play modules, the system performs both linear and nonlinear operations required for universal quantum computing. The study, published in Nature Photonics, marks a shift toward more practical, fault-tolerant photonic systems.
Traditional photonic quantum processors often struggle with universality because single-photon-level nonlinearities are difficult to achieve. Clavina adopts a design inspired by classical CPUs. A central control unit routes optical modes between a linear-optical network and specialized nonlinear modules. This approach allows the system to reuse hardware across different tasks, which significantly reduces the physical overhead typically associated with large-scale photonic processors. The team demonstrated this by using time-bin multiplexing to execute complex instructions on a single optical path.
Practical Applications and Non-Gaussian State Engineering
Beyond basic architecture, the team demonstrated the practical utility of Clavina through two specific applications. They generated optical Gottesman–Kitaev–Preskill (GKP) states, which are vital resources for bosonic quantum error correction. By applying real-time feedforward to interference patterns of Schrödinger cat states, the researchers produced these grid states quasi-deterministically. This process is a necessary step toward building fault-tolerant hardware that can protect information against noise.
Another application is the simulation of many-body quantum systems. Using the system's nonlinear unit, the team simulated a Bose–Hubbard model. This model describes how bosons hop between lattice sites and interact on-site. Unlike superconducting platforms, where on-site interactions are often fixed, Clavina allows for the tuning of these interactions on the fly. This flexibility enables the simulation of physics that were previously difficult to probe, particularly beyond the hard-core boson limit.
Path Toward Scalable Quantum Simulation
The Clavina architecture stands out because of its extensibility. Scientists can add new modules to the central control unit without redesigning the entire processor. This plug-and-play capability supports varied tasks, from quantum neural networks to Hamiltonian simulations. The team successfully used a dual-core linear-optical network to perform a 100-mode Gaussian boson sampling experiment, confirming the high phase stability of their platform over extended periods of operation.
While the current proof-of-concept uses bulk optics, the researchers state that the architecture is compatible with integrated photonic platforms like thin-film lithium niobate. Moving the design onto a chip could eliminate coupling losses and further increase the speed of operations. The researchers view this demonstration as a pathway to fault-tolerant universal photonic quantum computing, positioning light-based systems as prime candidates for future quantum hardware.

