The Financial Conduct Authority will convene banks, quantum computing providers and national research laboratories in a Quantum Experimentation Group, and its first job is financial crime. Participants will test whether quantum machine learning can detect money laundering and fraud, working from a synthetic dataset the regulator supplies.
What the FCA announced
The FCA published the plan on 5 November 2026. It calls the group the first of its kind for a regulator, and says the group will launch in the first quarter of 2027. Financial services firms, quantum technology providers and national research laboratories can register interest until 31 December 2026.
The FCA’s page for the group limits the invitation to UK organisations in three groups: firms that are only starting to look at quantum computing, firms that already run research and experiments, and the providers and labs that could help design and evaluate the tests. It describes the group as participant-led. Each participant stays responsible for building and assessing its own solution with its quantum specialists, and the FCA’s role is to define the problem and work alongside them.
Advertisement
Jessica Rusu, chief data, information and intelligence officer at the FCA, said: “We want to help firms explore where quantum could make a meaningful difference, and early experimentation is essential to understanding its potential in financial services.”
Why the dataset is the practical offer
The experiments need data, and most innovation teams cannot get realistic financial crime data. The FCA says so on its group page: “Suitable data also remains a practical barrier to experimentation.” Its answer is a synthetic dataset, which the press release describes as artificially generated data, shared with every participant.
The FCA’s October 2025 research note on quantum computing described the same constraint from the firm side. It reported that many quantum teams sit within innovation units with limited access to sensitive financial data, which forces reliance on open or synthetic datasets. A common dataset also gives the group a way to compare approaches. The group page says participants will use it to develop, test and evaluate quantum approaches, including comparisons against classical techniques.
That comparison is the part a compliance or fraud-operations lead can use. A quantum model that is tested only against itself says little. A quantum model scored against the classical detection a bank already runs, on the same data, produces a number that can go into a business case or be dropped from one.
What the FCA wants to learn about regulation
The FCA frames the group as a way to build its own understanding as much as the industry’s. The press release says it will help the regulator understand the technology’s opportunities and limitations, the barriers to experimentation and what these could mean for regulation. The group page lists developing early regulatory insights as one of four aims, alongside quantum machine learning for financial crime detection, building knowledge and institutional capability, and access to the synthetic dataset.
The October 2025 research note flags a regulatory question that this experiment can start to answer. Its section on validation, benchmarking and replicability says supervisors rarely expect perfect replication of models, but they do require clear benchmarks and evidence that outputs hold up and stay stable. For quantum systems, it adds, reproducibility will be complicated by the probabilistic nature of outputs and hardware variability, so firms may need to show that a quantum model’s output aligns with expected distributions. An AML or fraud model is exactly the kind of model a supervisor asks to see validated, which makes this use case a natural place to find out what that evidence looks like.
The FCA also states that participants’ commercially sensitive information will be protected. The group will share only high-level learning from the process.
What it means for the finance leader
Nothing in the announcement says quantum machine learning outperforms the models banks run today. The FCA’s wording is that the initiative aims to help firms understand where quantum computing might offer benefits over existing classical computing approaches. That is a test design, and the results are not in. No timeline for findings is given beyond the Q1 2027 launch.
For a head of financial crime or a chief risk officer, three things follow from what the FCA has published.
- The registration window closes on 31 December 2026, so a firm that wants a seat has about eight weeks to decide.
- Firms without quantum staff are explicitly in scope. The FCA says some firms have yet to develop the expertise or resources needed to begin, and the group is designed to connect them with providers and labs.
- The common dataset sets the terms. A firm that joins will be measured on the FCA’s problem statement, so the useful internal work before registering is to write down which of its own AML or fraud detection results the quantum approach would have to beat.
In our reading, the FCA is choosing to observe quantum through a controlled test before it writes any expectations. That sits beside the regulator’s separate and more urgent message on cryptography, covered in FCA Expects Firms to Prepare for Post-Quantum Cryptography Now. The two sit on different timelines. The cryptography work asks firms to act on a threat to their security, and the experimentation group asks them to evaluate a possible tool.
The financial crime problem the group takes on is large in operational terms. The FCA’s own survey, reported in FCA Survey: Firms Closed 238,396 Suspected Mule Accounts in 2025, shows the volume of suspected mule accounts firms are already closing. The group page does not say which metrics the experiments will use.
What to do before December
A firm weighing a place in the group can prepare without any quantum expertise. Pull the current false-positive and detection rates for the AML and fraud models in production. Name the person who would own the experiment, and decide in advance what result would justify a second round. The FCA says it will define problem statements and work alongside participants, so the firms that arrive with a baseline will get more from the cohort sessions than those that arrive with curiosity.
Interest can be registered through the FCA’s page for the group.
Source: FCA press release, “FCA to explore quantum computing to tackle financial crime”
