Xanadu

Xanadu is a Canadian quantum-computing company, based in Toronto, that pursues photonic quantum computing. It is known for the 2022 Borealis experiment, which reported a Quantum advantage result published in Nature, and for maintaining the open-source PennyLane software for quantum programming and quantum machine learning.

#Photonic approach

Xanadu builds its processors from photonic components, using squeezed states of light as the information carriers and integrated optical circuits to manipulate them. Photonic systems can operate much of their hardware without the millikelvin cryogenics that superconducting qubits require, and photons are natural carriers for networking between modules. The tradeoffs are the ones common to photonics: gates are effectively probabilistic and photon loss is a central obstacle, which shapes how the architecture must be error-corrected at scale. Xanadu made earlier photonic chips, such as an eight-mode device offered over the cloud, available to researchers before the larger Borealis experiment.

#Borealis and the 2022 advantage claim

In 2022 Xanadu reported a machine called Borealis that performed Gaussian boson sampling, a photonic sampling task, across 216 squeezed-light modes. The team argued that reproducing the sampled distribution would take a classical supercomputer far longer than the experiment took, and published the result as a demonstration of quantum computational advantage (Madsen et al. 2022). As with other sampling-based advantage claims, including Google's 2019 Quantum supremacy result, the task is contrived to be hard for classical machines and has no direct practical application, and the precise classical difficulty of such sampling problems has been probed and revised by later work. Borealis was a genuine milestone for photonic hardware, not a general-purpose or cryptographically relevant computation.

#PennyLane and software

Xanadu develops PennyLane, a widely used open-source library for differentiable quantum programming that connects quantum circuits with machine-learning frameworks and runs across multiple hardware backends. The software has an adoption footprint beyond Xanadu's own hardware, making it one of the company's more visible contributions to the field. As of early 2026, Xanadu has also described modular photonic systems aimed at networking many chips together as a route toward larger, fault-tolerant machines, which it presents as a goal rather than an achieved result.

#Status

Xanadu's demonstrated results are sampling experiments and software; it has not reported a fault-tolerant machine or an advantage on a practical problem. Its stated long-term direction, scaling and networking photonic modules with Quantum error correction, is an open engineering program.

#Relation to cryptography

Photonic machines can in principle be made universal and fault-tolerant, in which case a sufficiently large one could run Shor's algorithm. The Borealis advantage result concerns sampling and says nothing about breaking encryption. Xanadu's relevance to the Post-quantum cryptography threat model is prospective and depends on its scaling goals, which remain unproven as of early 2026.