Each year, fraud drains over $343 billion from merchants globally with 50% of consumers calling for their banks to have “better fraud detection systems” (Juniper Research 2022). This activity results in nearly 40% of cardholders abandoning their card and 20% of cards showing no activity for six months after a false decline (Fiserv, ABA Banking Journal). The impact on trust is immediate and there is a critical need for more accurate and robust fraud detection systems, with 71% of large financial institutions say they plan to start or improve their fraud solution (PYMNTS Intelligence 2024). 

The stakes are high with fraud detection systems needing to make decisions within milliseconds balancing precision, customer experience, and security at extraordinary scale.

As fraud tactics evolve and data grows more complex, classical computing will increasingly face constraints in both hardware and algorithmic capability. Financial institutions need smarter and more adaptive tools to stay ahead. To address this challenge, OQC and Mastercard have partnered on this project to explore quantum-enhanced fraud detection, combining advanced machine learning with OQC’s superconducting quantum hardware to push the boundaries of what’s possible in real-time financial security.

Why Quantum is Being Explored For Fraud Now
Fraud today is more dynamic, more automated, and more adversarial than ever before. Criminals continuously evolve their techniques, probing for vulnerabilities and exploiting weaknesses faster than rule-based or static models can adapt.

Classical machine learning systems face several well-known constraints:

• Highly imbalanced datasets – real fraud cases are rare, making training difficult
• Rapidly evolving fraud patterns – requiring constant retraining and feature engineering
• High false positives – where security measures inconvenience customers more than criminals

For this project, the objective is clear: reduce fraud without stopping legitimate customers. Quantum models can capture complex correlations in high-dimensional data, generalise better from limited examples, and are naturally suited to anomaly detection, where rare events hide within vast datasets.

Importantly, OQC’s superconducting qubits offer the fastest quantum modality available today, making them the best-positioned hardware platform for environments like payments.

Building a Practical Path to Quantum-Ready Fraud Systems
The project was designed around practical, commercially relevant goals. Together, OQC and Mastercard set out to:

• Evaluate how quantum algorithms can enhance fraud detection performance
• Benchmark quantum-enhanced models against Mastercard’s existing classical systems
• Build hybrid workflows that could eventually operate in real payment pipelines
• Explore operational constraints such as latency

This initiative focuses squarely on production readiness: proving where quantum could deliver real-world improvement.

OQC’s Hybrid Ensemble Architecture
To solve the latency challenge, OQC developed a hybrid system stack that strategically blends quantum and classical algorithms.

The system works by sending only the most complex and highest-value datapoints (those where quantum models provide the greatest improvement) to quantum hardware, while classical models evaluate the remaining datapoints, where classical performance is already strong and latency is critical. The outputs from both are then combined into a single fraud decision, optimised to deliver maximum accuracy without compromising speed.

This design allowed Mastercard to begin accessing the benefits of quantum machine learning without exceeding real-time latency constraints. It preserves the operational speed required for global payments while creating a practical path to explore deploying quantum in production systems. It also establishes the first realistic template for quantum-enabled fraud detection and provides a model that other financial institutions can adopt as they explore quantum technologies.

Early Results: Reducing False Positives
The early stages of the collaboration has shown a tangible reduction in false positives. By incorporating quantum in a mixture-of-experts ensemble with state of the art classical models, Mastercard and OQC demonstrated a reduction in false positives – the exact type of error that leads to customer frustration, lost revenue, and diminished trust.

Reducing false positives means:

• Fewer legitimate payments being incorrectly declined
• Stronger customer loyalty and retention
• Increased card usage and revenue
• Lower operational costs for dispute handling

Why This Work Matters for Financial Services
Quantum computing is developing at a time when financial institutions face unprecedented pressure from cybercrime, fraud, regulatory requirements, and customer expectations. To compete in this environment, the industry needs tools that can detect patterns traditional systems miss – tools that can scale with complexity and adapt quickly to emerging threats.

This project demonstrates that quantum technology can unlock new levels of performance in fraud detection, offering capabilities that extend beyond what classical models can achieve. By using hybrid approaches, it becomes possible to overcome real-world barriers such as latency, data movement, and system reliability, challenges that have historically limited quantum adoption in financial services. 

The collaboration also shows that quantum workflows can be designed to meet the stringent operational requirements of global financial institutions, proving that the technology is no longer purely theoretical but increasingly usable today. As a result, this project stands as one of the first meaningful demonstrations of how financial institutions can begin exploring how to integrate quantum computing into their cybersecurity and fraud prevention strategies.

Looking Ahead
With practical hybrid architectures and early evidence of reduced false positives, quantum computing is emerging as a serious tool for real-world financial systems. This work marks a pivotal step toward the long-term vision of secure, intelligent, quantum-enhanced payments and sets the direction for how fraud detection will evolve over the next decade.

Access the preprint here: https://arxiv.org/abs/2603.06473