I was invited to listen in on an event hosted by Fujitsu and Quantum Insider on the reality of Quantum Computing (QC) in financial services today. This market is a good test for QC since multiple possible high value applications have been suggested. The panel was chaired by Brian Lenehan, (Founder and Chair, Quantum Strategy Institute), with panelists Franco Severini (CTO Financial Services, Fujitsu), Philip Intallura (Head of Quantum, HSBC), Spencer Izard (Research Director, at industry analyst firm PAC) and Ellen Devereux (Quantum Computing Consultant, Fujitsu). Great opportunity to get past the hype and panelists did not disappoint – sharing guarded enthusiasm mixed with realism.
Quantum and finance – a deliberate, cautious approach
Our simplistic view of new technologies tends to binary choices: full-throttle adoption or wait-and-see, production-ready tomorrow or years in the future. Panel participants were much more thoughtful in balancing upside and downside risks, as you would expect in a domain where opinions can swing hundreds of billions of dollars.
Everyone agreed that there is significant promise in using quantum computers in finance for modeling probabilistic systems, for optimization and for quantum machine learning, and an eye on quantum risks around encryption which I won’t touch on here. There was equal agreement that hardware isn’t ready yet for production applications. What I hadn’t appreciated is the degree of preparation necessary in financial services to get ready for production.
Identifying use-cases for quantum starts with business priorities, not a library of standard quantum algorithms. Figuring out how quantum will fit into existing financial pipelines requires thought and planning. Staffing for expertise is another challenge. Where do you start – with QC experts or with quants (financial experts with deep mathematical and programming expertise)? Both face a steep learning curve, but the panel felt quants may be the better choice given their deep grounding in applications.
On this point it is important to understand a central challenge in applying QC in virtually any domain. What QCs do very well is to solve exponentially hard math problems in sub-exponential time. For anything else, control, data transport, even regular arithmetic, you’re much better off using a classical computer. Think of quantum as a specialized coprocessor attached to a classical computer. Algorithm development must start by figuring out an effective partitioning between these two systems. This task is non-trivial, even for an experienced programmer. Shor’s algorithm is a good example of this hybrid computation: mostly running in classical with a small (but exponentially hard) number theory problem running in QC.
Building teams of expert program developers will take time, and faces significant competition from other financial institutions. All panelists see accelerating wage inflation for such experts, not unlike competition for AI talent. I also sensed wide agreement, if not phrased this way, that QC in financial services starts by design with bespoke partner support. Fujitsu for example offers such services to build momentum around QC while learning about and preparing for high value opportunities, while also remaining nimble as tech and market needs evolve.
The point of QC is algorithm acceleration, but the level of acceleration that will be interesting is highly dependent on the application. One panelist said that quantum Monte Carlo for high dimensional derivative pricing today only shows quadratic advantage, which for him is not interesting. He needs exponential advantage to keep up with exploding problem sizes. In contrast, for applications supporting high volume transactional processes such as fraud detection, almost any level of performance improvement is interesting.
Overall, the view was that it will take 5 years to prepare good understanding of best targets, to optimize (hybrid) algorithms to meet those targets, and to staff up with experts able to manage technical algorithm complexities. Enough time perhaps for physical QCs to become ready.
For hype enthusiasts claiming 2026 QC is the year of quantum, yes financial services are making “measured investments” to get ready and are working with partners like Fujitsu to help develop expertise. But production deployment is generally agreed to be 5 years out and still cautiously funded.
Fujitsu and QC
I haven’t covered Fujitsu so far in my quantum series, but I now know that they are a serious contender and have been for some time. Their production QC supports 256 physical qubits. A 1,000-qubit system is planned for later this year with QEC (quantum error correction) sounding somewhat like the IBM approach. Users can access the technology through a hybrid quantum platform, supported by libraries to simplify development, also through classical-hosted quantum simulators.
Fujitsu is also collaborating with a company building on diamond spin qubits and reporting fidelity levels (difference between a noisy quantum state and the ideal state) among the highest today. Good for Fujitsu in spreading their bets, given proliferation of QC technologies. They have also announced interesting joint development with Osaka university which should allow them to dramatically reduce the number of noisy qubits required in certain calculations, extending the value of QC in the NISQ (noisy intermediate scale quantum) era.
I’m adding Fujitsu Quantum to my list of Quantum enterprises to watch.
Also Read:
An Upper Bound on Effective Quantum Computation?
Another Quantum Topic: Quantum Communication
PQShield on Preparing for Q-Day
Share this post via:
Intel, Musk, and the Tweet That Launched a 1000 Ships on a Becalmed Sea