Xanadu Quantum Technologies

Overview

Xanadu is a Toronto company that bet early on photonics: building quantum computers from light rather than from superconducting circuits or trapped atoms. Photons travel through optical fiber and silicon chips, interact weakly with their environment, and can be manipulated at or near room temperature, although the detectors that read them out still need cooling. That makes photonic designs attractive for networking many modules together, which is central to Xanadu's strategy.

The company is also known for PennyLane, an open-source software library for quantum machine learning and differentiable quantum programming that is used by researchers across the industry whatever hardware they run on. That software reach gives Xanadu a developer audience far larger than its hardware customer base.

History and background

Christian Weedbrook, a physicist with a background in continuous-variable quantum information, founded Xanadu in 2016 after writing a whitepaper arguing that photonics offered a path to scale. The company released the Strawberry Fields software library, then PennyLane in 2018, and in 2022 demonstrated a machine called Borealis that sampled from a quantum state of light in a way Xanadu said would be impractical to simulate classically. Borealis was made available through the cloud.

The company's next phase focused on fault tolerance. In 2025 it introduced Aurora, which it describes as the first modular, networked photonic quantum computer with real-time error correction, and it reported 12 logical qubits encoded in a bosonic scheme called GKP. On March 27, 2026 Xanadu began trading on Nasdaq and the Toronto Stock Exchange under the ticker XNDU after completing a merger with the special purpose acquisition company Crane Harbor Acquisition Corp.

How the technology works

Most photonic approaches encode quantum information in features of light such as its path, its polarization, or its squeezing. Xanadu works with continuous-variable states, in which information is stored in the amplitude and phase of the light field. A particular encoding called the Gottesman-Kitaev-Preskill (GKP) state packs a qubit into a single mode of light in a way that makes small shifts detectable and correctable. Generating high-quality GKP states is one of the hardest problems in the field, and Xanadu's reported 12 logical GKP qubits with real-time error-correction decoding is a published step in that direction.

The central enemy for photonic machines is loss. Every optical component absorbs or scatters a tiny fraction of the light, and those losses add up as circuits grow. Xanadu cites a roughly twenty-fold improvement in the performance of its photonic components over the past three years, and it has been working with semiconductor partners such as Tower Semiconductor and ASML to build its chips on production fabrication lines. It has also announced a partnership with the cryogenics company Bluefors.

Aurora connects multiple rack-mounted modules with optical fiber, and the company's roadmap expects further modules to be networked as the system scales. At its August 2026 roadmap presentation, Xanadu described a plan to reach fault-tolerant operation in 2028 to 2029, then to build a quantum data center, and to reach more than 1,000 logical qubits by 2031. Earlier statements had set a goal of up to 500 logical qubits in 2029 to 2030. These are targets, and the figures have moved between presentations.

Products and platforms

Processors in our Hardware Database

• Xanadu Borealis — 216 qubits, released 2022

Key milestones

Strengths and challenges

Strengths

• A scalable architecture based on networking modules over optical fiber.
• PennyLane gives the company a large developer community and influence beyond its own hardware.
• Peer-reviewed results on logical GKP qubits and a modular design.
• Partnerships with chip fabrication and cryogenics leaders, and government support from the United States and Canada.

Challenges and open questions

• Optical loss still limits circuit depth, and further large reductions are needed.
• Generating GKP states at high quality and high rates is extremely difficult.
• Xanadu describes itself as in an investment phase, so it is spending heavily and revenue is not the main story yet.
• Roadmap targets have shifted between presentations, so the 2028 to 2029 fault-tolerance goal should be read as a goal.

Why it matters

Photonics is the most radical of the major quantum hardware approaches, and Xanadu is its best-known public company. If photons can be made low-loss and generated on demand, quantum computers could be built with the manufacturing methods and networking that already support the telecom and chip industries. If not, the approach could remain a niche.

For students, PennyLane is also a practical entry point. You can write a quantum machine learning model in Python and run it on simulators or on hardware from several providers, which makes Xanadu's software relevant even if you never use its computers.

Latest news

• Xanadu presents a roadmap to more than 1,000 logical qubits by 2031 and fault-tolerant operation in 2028 to 2029 — August 2026
• Xanadu announces plans for an advanced photonics research and manufacturing facility — August 2026
• Xanadu reports Q2 2026 results with 312.8 million dollars in cash and raises 67.2 million dollars through an at-the-market facility — August 2026
• Xanadu and ASML announce a collaboration on advanced lithography for photonic chips — 2026

Frequently asked questions

This profile is written for educational purposes and summarizes public announcements, company filings, and published research. Last reviewed: September 30, 2026. It is not investment advice. Check the company's official channels for the latest figures and claims.

← Back to all quantum companies