Photonic

Xanadu

Xanadu, based in Toronto, builds photonic quantum computers that encode information in pulses of light travelling through chips and optical fibre. It uses an error-resistant way of encoding light (called GKP states) and aims to link many identical rack-sized modules by fibre into one larger machine. Much of the hardware works at room temperature, though some subsystems may need temperature control. Xanadu also develops the open-source PennyLane software. Its 2025 Aurora system is a small demonstrator, not a commercial product.

publicly listed company (Xanadu Quantum Technologies Limited, listed on Nasdaq and the Toronto Stock Exchange as XNDU after its SPAC merger with Crane Harbor Acquisition Corp. in March 2026) · Based in Canada · Since 2016 · Facts checked 27 October 2026

Where it ranks this month

Best place in each race, from QubitBenchmark release 2026-11.1.

All rankings →

Most accurate–No ranked result Error-corrected qubits–No ranked result Most qubits–No ranked result Fastest–No ranked result Use it today–No ranked result

Machines

Machine	Physical qubits	Introduced	How to use it
Aurora (12-qubit modular photonic demonstrator, 35 chips, 4 racks)	12	2025	not publicly available (company says no systems are for commercial sale beyond beta stage)
Borealis (Gaussian boson sampling device, 216 squeezed-light modes)	216	2022	retired from Amazon Braket in June 2023

Qubit counts as the company reports them. Our rankings use QubitBenchmark's checked results instead.

Promised vs delivered

What Xanadu has said it would build, and what happened.

All roadmaps →

Milestone	Target	Status	Why we say so
Modular, networked photonic computer (Aurora)	No date given	Delivered	Aurora was described in a peer-reviewed Nature paper published 22 January 2025, which itself calls it a scale model and says optical loss remains the main hurdle to fault tolerance. Source
Device performance improvement and early fault-tolerance building blocks	through fiscal 2026	Due this year	Q2 2026 results report an edge-coupling loss of 0.085 dB per facet, but the August 2026 roadmap still puts photon loss at 24.1x above the fault-tolerance threshold in 2026. Source
Qubit Factory build	2026-2027	Unclear	Listed in the August 2026 roadmap; no progress report found. Source
Fully fault-tolerant quantum computer	'over the next three years' (November 2025); 2028-2029 (August 2026)	Unclear	The CEO said in August 2026 that loss reductions give confidence in reaching fault tolerance in 2028 to 2029; no independent evidence yet. Source
Quantum data centre (in Canada)	2029 (November 2025); 2029-2030 (August 2026)	Slipped	The November 2025 target of 2029 became a 2029-2030 window in the August 2026 roadmap. Source
Up to 100,000 physical qubits and up to 500 logical qubits	2029-2030	Unclear	Stated in the annual report, which also says only twelve logical qubits have been shown so far, with photon loss not at required levels. Source
Up to 200 logical qubits (2029), up to 500 (2030), 1,000+ (2031)	2029 / 2030 / 2031	Unclear	Company roadmap targets presented August 2026; not yet testable. Source

Recent claims, and how far the evidence has got

Peer-reviewed Aurora: 12-qubit modular photonic computer from 35 chips networked by fibre, with a cluster state of 86.4 billion modes and a distance-2 repetition code decoded in real time Source
Peer-reviewed First on-chip generation of error-resistant GKP photonic qubit states (silicon nitride on 300 mm wafers) Source
Announced Average fibre-to-chip edge-coupling loss of 0.085 dB per facet Source

In other organisations' rankings

Where Xanadu appears in other published rankings and lists. They don't change ours.

All of them →

The Quantum Index Report 2026, MIT Initiative on the Digital Economy (MIT Sloan), in collaboration with Accenture (2026-10). Transparent, sourced and openly licensed data, but it measures activity (patents, funding, press releases) rather than how well the machines perform.
CB Insights ESP matrices: Quantum computer makers (neutral atoms, photonics, silicon qubit, superconducting circuits, trapped ions), CB Insights (2026). Based on market and business signals, not machine performance. Vendors can brief CB Insights, and most placements are hidden behind a subscription.
Quantum Benchmarking Initiative: Stage B Selection, DARPA (US Defense Advanced Research Projects Agency) (2025-11-06). An official list of who advanced, not a ranking; it partly rests on plans and roadmaps, and Stage C has since narrowed it (see darpa-qbi-stage-c-2026).

Sources

Xanadu introduces Aurora: world's first scalable, networked and modular quantum computer, Xanadu via Cision Newswire, 22 January 2025
Scaling and networking a modular photonic quantum computer, Nature, 22 January 2025
Xanadu Demonstrates Scalable Building Block for Photonic Quantum Computers, The Quantum Insider, 5 June 2025
Xanadu to go public on Nasdaq and TSX in $3.6-billion USD SPAC deal, BetaKit, 3 November 2025
Xanadu Quantum SPAC Deal Approved, Targets $302 Million Raise, The Quantum Insider, 19 March 2026
Canadian quantum firm Xanadu debuts on public markets after SPAC deal, Yahoo Finance, 30 March 2026
Xanadu Quantum Technologies Ltd, Form 20-F for fiscal year 2025, U.S. SEC
Xanadu Reports Higher Revenue in 2Q 2026, Expands U.S. Operations, The Quantum Insider, 6 August 2026
Xanadu Targets More Than 1,000 Logical Qubits by 2031, The Quantum Insider, 31 August 2026
Xanadu unveils technical roadmap targeting 1,000 logical qubits by 2031, Quantum Computing Report, 31 August 2026
Rigetti's 84-qubit Ankaa-2 quantum computer available through AWS Braket service, SDxCentral, 27 August 2024