Fujitsu and Daiichi Put Quantum Asset Management on Trial

Fujitsu and Daiichi Life Group are testing quantum asset management on a 30 trillion yen insurance portfolio. The yearlong research will build and benchmark algorithms for asset allocation across stocks, bonds and alternative assets, using simulators and quantum computers while factoring risk-return targets, liabilities, regulation and investment constraints.

The companies announced the project on June 4, with research running from April 2026 through the end of March 2027. The release prices the experiment in basis points: a one-basis-point improvement on the portfolio would add 3 billion yen of return before fees, taxes or implementation costs.

The One-Basis-Point Price Tag

A basis point is one hundredth of a percentage point. On 30 trillion yen, about $188 billion, Fujitsu and Daiichi Life Group say that small move equals 3 billion yen. That is the number that turns quantum asset management from a lab phrase into a balance-sheet question.

The research target is asset allocation, the routine decision that becomes difficult when the number of assets, scenarios and restrictions grows. Daiichi Life Insurance has to weigh stocks, bonds and alternative assets against future insurance liabilities, capital rules and limits on what each asset class can hold.

Fujitsu and Daiichi Life Group are saying the first gain they want to test is better search. A portfolio team can model many combinations today, but each added constraint narrows the space that a usable answer has to satisfy. The companies plan to evaluate allocation patterns under a wider range of economic scenarios than a simple risk-return model would cover.

Fujitsu Brings the Simulator to the Insurance Desk

Fujitsu enters the project with a quantum stack already aimed at corporate application research. Its 40-qubit quantum simulator overview describes a CPU, or central processing unit, based state-vector simulator built on 1,024 FX700 nodes powered by A64FX processors. The same overview says Fujitsu enhanced Qulacs, an open-source quantum circuit simulator, for parallel execution across the cluster.

The Daiichi work starts before fault-tolerant quantum machines exist at financial scale. That is why the simulator sits at the center of the plan. Fujitsu will provide quantum algorithm expertise and computing environments, while the tests will also use available quantum computers for performance checks.

Fujitsu has been building the hardware side in public. In April 2025, Fujitsu and RIKEN, Japan’s national research institute, said in their 256-qubit quantum computer announcement that the superconducting machine would be added to a hybrid quantum computing platform for outside users. Fujitsu’s own quantum material also points to finance as one of the intended industrial uses, along with materials and drug discovery.

Why an Insurer Makes a Demanding Test Case

Life insurers invest against promises that stretch for decades. Daiichi Life Group’s fiscal 2026 results filing listed 74.159 trillion yen of total assets at March 31, 2026, while the June 4 research release uses the 30 trillion yen managed-assets figure for Daiichi Life Insurance’s operations.

Asset liability management (ALM, matching invested assets to future policy obligations and capital rules) is where quantum research enters the insurance workflow. A Society of Actuaries Research Institute report said strategic asset allocation and ALM may benefit from quantum computing because the number of possible asset allocation plans can become too large to test in a practical time frame.

The constraints Daiichi brings are ordinary insurance constraints, which makes them useful for testing:

• Future benefit payments and surrender behavior have to be reflected in the liability profile.
• Interest-rate, credit and equity shocks change both asset values and capital needs.
• Regulatory limits can rule out an allocation even when a return model likes it.
• Alternative assets add liquidity limits and valuation lag to the search problem.

Daiichi Life Group will define research themes, evaluation metrics and the business data used in the test. The company will also supply workflows and on-site problems from its asset management operations, which gives Fujitsu’s algorithms a business setting beyond a synthetic portfolio.

The Work Is Split Between Data and Machines

The project runs for one year, and the roles are narrow enough to keep the research measurable. Fujitsu is the quantum technology supplier. Daiichi Life Group is the insurance operator whose data and constraints shape the test. Both companies will design and develop the algorithms.

The planned verification will vary constraints and simulation scenarios, then compare performance on simulators and quantum computers. The release names verification, scenario testing and algorithm design as the year’s outputs, with business deployment tied to future large-scale, high-performance quantum computers.

The project also gives Daiichi a way to test data readiness. Quantum algorithms still need clean inputs, clear constraints and baselines from existing systems. A poor allocation model remains poor when it is run on a more exotic machine.

Quantum Finance Has a Queue Already

Daiichi’s project lands in a busy corner of financial technology. The European Securities and Markets Authority (ESMA, the European Union securities markets regulator) said in its quantum financial markets analysis that financial use cases include portfolio optimization, transaction settlement, Monte Carlo simulations, fraud detection and credit scoring. ESMA also said current quantum hardware remains insufficient to solve industry-scale financial problems more efficiently or accurately than established classical approaches.

Asset managers have been testing the same idea from another direction. IBM and Vanguard said in their portfolio optimization research note that they explored variational quantum algorithms for portfolio construction under real-world constraints, using a hybrid workflow that combined quantum sampling with classical post-processing.

The Daiichi project is narrower than a broad finance platform. It is tied to one insurer’s investment process and one year’s verification schedule. That gives the companies a clearer comparison set: existing allocation tools, internal risk metrics and the portfolio rules Daiichi already uses.

The Benchmark Will Decide What Survives

The useful output from the year will be comparative: solution quality, run time and stability against the classical solvers Daiichi already trusts. Daiichi needs an answer that survives model review, audit trails and the committees that approve changes to strategic allocation.

The scorecard will likely come down to practical items:

• Optimality gap versus current classical methods.
• Run time under larger scenario sets.
• Stability when constraints change.
• Reproducibility of results on the simulator and available quantum computers.
• Data handling controls for sensitive insurance portfolios.

Financial institutions are slow to move allocation tools into production because a model can change capital usage, hedge demand and the timing of trades. A portfolio tool that changes allocations would need governance before it touched policyholder assets.

The release also says the companies plan to share insights through academic papers and other publications. That matters for outside readers because a paper can show the benchmark setup, the classical baseline and the limits of the quantum method. Without those details, the work remains a corporate proof of concept.

March 2027 Is the First Deadline

The partners started the research in April 2026 and plan to finish by the end of March 2027. By then, Fujitsu and Daiichi Life Group should know which parts of the allocation problem can be mapped cleanly to quantum algorithms and which parts still belong with existing systems.

A result by March 2027 would arrive while Fujitsu is expanding its hardware platform and while regulators are still describing quantum finance as an early-stage field. The public signal will be the detail in papers: benchmark data, constraint sets and baseline methods.

The research period runs through the end of March 2027.