By Dr. Raju Gudla, Assistant Professor, Department of Computer Science and Engineering, SRM University -AP (Amaravati)
Financial services firms are looking at quantum computing as a possible technology for addressing the computationally demanding optimization, risk management, and fraud detection problems. In Banking, financial services, and insurance (BFSI) applications, there tend to be many variables, constraints, uncertainty, and data that are fast-changing. The classical computing approach continues to dominate in these applications; however, quantum computing presents possible alternate technologies to address some optimization and machine learning problems. Advances in quantum software and quantum applications have made it possible for researchers in finance to try out these features on actual quantum computers.
Quantum Computing and the Financial Sector:
Many problems in finance have their roots in mathematical optimization. Portfolio construction entails choosing financial instruments taking into account expected returns, risks, diversification, transaction costs, liquidity, and restrictions on investments. With increasing numbers of assets and restrictions, the size of the search space may become enormous. Not all financial workloads can be solved using quantum computing. The latter’s strength lies in solving particular computation architectures which can be formulated as either optimization or sampling problems. That is why the current practice utilizes hybrid quantum-classical workflows in which classical computing systems perform preparation of financial data and formulation of the optimization problem, while quantum computers are applied for certain computation tasks. Examples of recent progress by IBM Quantum and other companies demonstrate the shift from theoretical research to experimental application development.
Quantum Portfolio Optimization:
Portfolio optimization is one of the most studied financial applications in the context of quantum computing. In a classical scenario, the problem of portfolio optimization consists in maximizing the expected returns under control of risks within the portfolio. With other factors, like transaction costs, limitations on assets, need for diversification and rebalancing, being taken into account, the problem may become complicated from the point of view of combinatorial optimization. The Qiskit Function Catalog from IBM at present features the Quantum Portfolio Optimizer created by Global Data Quantum. This function deals with dynamic portfolio optimization and employs the approach based on the Variational Quantum Eigensolver (VQE) to solve the problem and represent it as the problem of Quadratic Unconstrained Binary Optimization (QUBO). The function allows obtaining the optimal trajectories of investments as well as some financial metrics, like the expected return and Sharpe ratio.
Another key element in this process is the ability to evaluate the performance of the strategy using historical financial data. Instead of considering quantum optimization as only a purely theoretical tool, it becomes possible to test how the various investment strategies might perform in a specified time period in the past. IBM notes that the Quantum Portfolio Optimizer was tested on various categories of assets, where the primary evaluation criteria were objective cost and the Sharpe ratio. Another improvement comes from Iskay Quantum Optimizer developed by Kipu Quantum. The optimizer is available through the Qiskit Functions Catalog of IBM and offers binary optimization without constraints in QUBO and higher-order HUBO formulations. IBM mentions the one-to-one mapping between binary variables and qubits in the optimizer, while Kipu provides applications for portfolio selection on a 250-asset universe of S&P 500. Kipu also provides benchmarking results comparing its approach to quantum optimization with the classical one. This information should be understood in the context of the particular benchmark, hardware, solver configuration, and evaluation criteria used rather than considered as proof of the universal superiority of quantum computers in financial optimization.
Quantum Approaches to Risk Modelling:
Another potentially important area of application is that of risk management. Financial organizations constantly estimate their exposure to risks including market risks, credit risks, liquidity risks, and risks in their portfolios, and where computations often require numerous simulations involving uncertainties in financial variables. Quantum algorithms have been devised for solving problems in Monte Carlo simulation, probability evaluation, derivative pricing, and risk assessment. The attraction is especially great for situations in which a financial organization needs to analyse many different future scenarios. Modern reviews of the field point to portfolio optimization, derivative pricing, and risk management as the main fields within quantum-finance research. However, most practical applications are still in the research and experimental stages, and achieving quantum advantage depends on various considerations such as choice of algorithm, size of quantum hardware, error rates, data loading costs, and availability of sufficiently large quantum computers. The best model, then, for BFSI organizations would be a hybrid one. They could use classical systems to do all data preparation and statistical estimations, and quantum computing would be used to address some hard problems.
Quantum Computing for Fraud Detection:
Detection of fraud is inherently a task of classification and anomaly detection. Financial organizations process financial transactions considering the features like transaction amount, frequency, merchant details, geo-behaviour, account history, and relationships between accounts. Machine learning algorithms based on the principles of quantum computing can contribute to the development of new methods for classification, feature representation, optimization, and anomaly detection. One promising direction might be building hybrid models of quantum machine learning, where classical machine learning algorithms or networks are combined with quantum optimization or feature processing algorithms. Recent studies have considered hybrid methods for financial fraud detection. For example, a 2025 study suggested a hybrid deep belief network enhanced with quantum optimization for detecting financial fraud. This indicates a growing trend in combining the techniques of quantum computing with traditional machine learning algorithms. However, this does not mean that quantum computing can replace existing algorithms of fraud detection. The use of quantum optimization algorithms can help solve hard subtasks of fraud detection, including feature selection, model parameter optimization, transaction clustering, or rule optimization.
From Proof-of-Concept to Financial Deployment:
The critical question for BFSI industry goes way beyond whether or not the quantum algorithm can solve a finance problem. The critical issue lies in whether the algorithm can add any value to that particular problem compared to the best classical counterpart.
Therefore, future evaluations should consider several dimensions:
- solution quality;
- execution time;
- number of quantum resources required;
- classical preprocessing and post-processing overhead;
- financial metrics such as Sharpe ratio and expected return;
- robustness to hardware noise;
- scalability with increasing numbers of assets or transactions;
- and reproducibility across different datasets.
One such example of practical application of the technology is Qiskit Functions from IBM. The repository contains industry-specific quantum algorithms, like portfolio optimization and general optimization algorithms, which help researchers interact with quantum computing hardware without implementing the full flow of hardware process.
Future Outlook:
In the near term, the path for quantum computing in the finance industry will be dominated by hybrid quantum-classical solutions rather than replacing classical computing infrastructure altogether. Financial companies have already developed highly optimized classical infrastructures, huge historical datasets, and machine learning workflows. Therefore, quantum technologies have to prove their superiority on particular computational challenges from an economic point of view.The optimization of portfolios is among the most realizable use cases today, as financial goals may be expressed via the formulation of a QUBO or higher order optimization problem. Risk modelling and fraud detection are some other areas to consider but require additional progress in developing quantum algorithms, encoding data, mitigating errors, and validating on the large scale.In the end, quantum computing might become another computational tool within the financial technology ecosystem. Portfolio managers could use quantum optimization for analysing investment strategies under constraints, risk teams could analyse quantum simulation techniques, and fraud analysis systems could leverage hybrid quantum-classical approaches for solving high-dimensional problems.According to the currently available literature, one can formulate a fairly confident but still moderate conclusion that quantum computing moves from purely theoretical applications in finance toward experimental optimizations and machine learning approaches, although the quantum advantage for BFSI still hasn’t been established. Further benchmarking against classical approaches, realistic data from financial markets, and actual deployment problems will decide on which applications will move from theory to practice.Against this background, the development of such services as IBM Quantum Portfolio Optimizer and Iskay from Kipu Quantum represents a significant step: instead of developing a full workflow with quantum computing from scratch, financial companies get an opportunity to try quantum optimization through application interface.
