OQC, Citi, and NQCC Evaluate Quantum-Compressed PINNs for Financial Derivative Pricing Workflows

OQC, Citi, and NQCC Evaluate Quantum-Compressed PINNs for Financial Derivative Pricing Workflows Quantum platform developer Oxford Quantum Circuits (OQC) has completed a joint technical case study in collaboration with global financial institution Citi and the UK’s National Quantum Computing Centre (NQCC). Conducted under the NQCC’s SparQ user-engagement initiative, the research evaluated the operational viability of combining parameter-efficient quantum machine learning with physics-informed neural networks (PINNs) to solve partial differential equations (PDEs) governing financial derivative pricing. Derivative pricing engines rely on partial differential equations—such as the Black-Scholes, Dupire local-volatility, and Hull-White framework models—to construct continuous option-pricing surfaces and compute risk sensitivity measures (Greeks). The joint initiative implemented Quantum-compressed Physics-Informed Neural Networks (QPINNs) using the Quantum-Train compression protocol. By embedding parameterized quantum circuits into shallow neural network layers, the model learns the continuous solution to governing pricing equations while reducing classical parameter footprints. [ OQC, Citi & NQCC Hybrid QPINN Derivative Pricing Parameters ]Modeling ComponentAlgorithmic & Architectural DesignMeasured Performance Trade-OffsFinancial Problem Domain• PDE-Governed Derivative Pricing Surfaces• Black-Scholes & Local Volatility Models• High-dimensional option payoff evaluation• Continuous pricing surface extrapolationQuantum-AI Compression Stack• Physics-Informed Neural Networks (PINNs)• Quantum-Train Parameterized Circuit Layer• Substantial parameter count reduction• Pricing errors maintained within classical orderOperational Bottlenecks• Classical PINN Baselines vs. QPINNs• Gradient-based loss function minimization• Pricing kinks/discontinuities increase error• Quantum circuit sampling latencies on current QPUs The benchmarking results demonstrated that applying quantum circuit representations to key hidden layers substantially compressed model complexity without degrading pricing accuracy beyond acceptable baseline orders of magnitude. However, the study identified key operational constraints: compression efficacy varied depending on layer placement, and areas with sharp pricing discontinuities or high volatility sensitivities introduced higher modeling error rates. While QPINNs offer significant parameter reduction, classical execution overheads—including repeated quantum circuit evaluation and gradient backpropagation—highlight the necessity for deeper classical-quantum integration as hardware matures. The project aligns with OQC’s application-focused strategy to develop co-designed quantum workflows for enterprise markets. The case study builds upon OQC’s broader benchmarking efforts across financial risk pipelines, complementing its recent joint algorithmic resource study with Trust Base on Credit Valuation Adjustment (CVA) and Value-at-Risk (VaR) modeling. Review the case study via Oxford Quantum Circuits here, examine program details on NQCC SparQ here, and read our prior technical analysis of OQC and Trust Base’s Benchmark of Hybrid Quantum Financial Workloads here. September 22, 2026 Mohamed Abdel-Kareem2026-09-22T18:13:04-07:00 Leave A Comment Cancel replyComment Type in the text displayed above Δ This site uses Akismet to reduce spam. Learn how your comment data is processed.