Xanadu Quantum Technologies reported higher year-on-year second-quarter revenue on August 5, 2026 as the Toronto-based photonic quantum computing company continues to trade near-term losses for long-run capacity. Q2 revenue reached $1.5 million, up from $1.1 million a year earlier, driven by DARPA program work. Xanadu ended the quarter with $312.8 million of cash after tapping $67.2 million from its Yorkville Advisors synthetic ATM facility.
Photonic chips: a new benchmark
The company said it hit an average edge-coupling loss of 0.085 dB per facet, meaning more than 98% of the optical signal survives each fiber-to-chip interface — a metric Xanadu describes as an industry benchmark. Foundry throughput for its thin-film lithium niobate and silicon nitride photonic chip platforms rose roughly 75% and 50% respectively, alongside expanded packaging work with DARPA, Corning and DISCO.
Software wins and enterprise pull
Xanadu released updated versions of the PennyLane and Catalyst quantum programming platforms, cut Toffoli-gate operations for Quantum Read-Only Memory by roughly half and trained a Fourier-based quantum machine-learning model with more than one million parameters on ribosomal RNA distributions. The company deepened partnerships with Oak Ridge National Laboratory — bringing PennyLane onto the Frontier exascale supercomputer — renewed its multi-year computational fluid dynamics program with Rolls-Royce and expanded a workforce initiative with Lockheed Martin.
Betting on Albany
Xanadu also confirmed a significant U.S. expansion anchored on Albany, New York, where its workforce has grown more than fivefold since 2023 and additional hiring is planned before year end. The scale-up mirrors moves by U.S.-listed quantum peers — IonQ has committed $15 million to Chattanooga (see TRM's coverage of the Tennessee Quantum Communications Research Center) while Japan's OptQC and Fujitsu logged capital and research alliances in the same week.
Reporting based on coverage from The Quantum Insider, GlobeNewswire and SEC filings.