How HSBC and Haiqu Unlock Scalable Quantum Risk Models

A collaboration between HSBC and quantum software startup Haiqu could address one of the most persistent challenges in quantum computing for finance.

The work centres on overcoming the data bottleneck that has prevented quantum computers from processing complex financial information.

The research demonstrates that quantum systems can now handle the probability distributions used in risk modelling.

This could mean practical applications are closer than the industry previously estimated.

Solving the loading problem

According to the firms' press release, financial institutions can provide data to quantum computers through a process called Quantum State Preparation.

The challenge lies in encoding heavy-tailed distributions, which are mathematical models that predict extreme market crashes.

Traditional encoding methods require complex circuits that current quantum hardware cannot process. These circuits overwhelm the system and cause calculations to fail before completion.HSBC and Haiqu used a method called Matrix Product States to create shallow circuits.

This approach provides a more efficient way to pack data into quantum systems.

The sampling-based workflow "avoids storing the full discretised dataset in classical memory, enabling larger encoding circuits to be generated," the firm says.

This removes the need to store every data point in memory simultaneously.

Real-world testing on IBM hardware

The teams conducted tests on IBM quantum computing processors including Eagle and Osprey.

These systems are designed to handle complex workloads at scale.

At 25 qubits, where physical quantum processors begin to handle complex data, the team used IBM hardware to reproduce probability distributions. The results satisfied all standard statistical benchmarks.

The tests expanded to 64 qubits to assess how the system performs under realistic noise conditions.

The team implemented a workflow to run circuits on larger processors with inherent imperfections.

Simulations at 156 qubits demonstrated the method could scale to manage datasets far larger than previous approaches allowed. This could show the technique has potential beyond current hardware limitations.

“Preparing complex probability distributions efficiently is a key step in many quantum algorithms,” explains Dr Philip Intallura, Group Head of Quantum Technologies at HSBC.

“This work shows how they can be implemented with much shallower quantum circuits, bringing practical applications such as financial risk modelling closer.”

Why this matters for your wallet

Risk modelling helps banks calculate potential losses during market downturns.

Accurate models allow institutions to maintain stability and protect customer assets.

Mykola Maksymenko, Co-founder and CTO of Haiqu, says: “One of the biggest practical barriers is getting realistic financial data onto today’s quantum hardware.

“This work shows a scalable path around that barrier and helps move quantum finance workflows from theory toward execution.”

The research could indicate that quantum technology for financial applications is approaching practical implementation.

While quantum computers are not yet deployed in retail banking operations, the technical barriers to their use appear to be diminishing.