Britain’s quantum funding programme is turning towards the engineering required to make quantum processors useful inside existing computing environments, asking researchers to demonstrate how the technology can work alongside high-performance computing and artificial intelligence rather than as an isolated machine.

UK Research and Innovation has opened an EPSRC funding competition for collaborative projects that integrate quantum computing into high-performance computational workflows. Applications opened on 4 September and close on 3 November, with projects expected to start on 1 June 2027 and run for up to three years.

The full economic cost of an individual project can reach £3.75 million, of which EPSRC will fund 80%. Quantum computing must remain central to the research, while proposals have to address a defined application and involve an identified end-user rather than producing a general technical demonstration with no operational customer.

The structure addresses a problem the quantum industry will face even if hardware performance improves quickly. Organisations already run scientific and industrial workloads across CPUs, GPUs, accelerators, cloud platforms, data pipelines, and mature software libraries, so a useful quantum processor will have to fit into that environment rather than expect users to rebuild every application around a separate machine.

Integration starts before hardware is mature

UKRI identifies several technical bottlenecks between a quantum processor and an end-to-end workflow. They include hardware descriptors that help software decide where a job should run, performance models for allocating resources, efficient data encoding, transpilation, error mitigation, and orchestration across classical and quantum systems.

These are systems-engineering problems rather than questions about quantum physics alone. An algorithm may offer a theoretical advantage, but an application can lose it if data preparation, network latency, queueing, error handling, or movement between machines consumes more time and energy than the quantum calculation saves.

The funding call therefore asks projects to build demonstrators showing where complete hybrid workflows provide meaningful advantages. Applications can cover chemistry, materials, optimisation, digital twins, and AI for science, while researchers are expected to measure practical outcomes including time to solution, fidelity, cost, energy efficiency, and usability.

Comparison with classical systems is important because conventional computing is not standing still. GPUs, specialised accelerators, improved algorithms, and AI-assisted scientific computing can all change the baseline during a multi-year quantum project. Claims of advantage need to be tested against a strong practical alternative rather than an outdated reference workload.

End-users enter the architecture discussion

EPSRC expects end-users to help define requirements and validate the resulting systems using realistic data where appropriate. Potential areas include drug discovery, energy infrastructure, financial risk, transport optimisation, manufacturing, materials, healthcare, and defence, with participants able to come from industry, government, public bodies, or the third sector.

That requirement changes the incentives compared with hardware-led research. A manufacturer interested in production scheduling will care about whether a hybrid system connects to existing planning software and returns results within an operational decision window, while a scientific team may value fidelity and throughput more than an isolated measure of quantum operations.

Collaboration with the National Quantum Computing Centre is optional, although successful projects may be able to use its hardware, simulators, facilities, and technical expertise. The wider Quantum Programme is also expected to attract at least one pound of public or private leverage for each pound of programme funding over its lifetime.

The call sits within the UK’s wider quantum strategy, but it deals with an integration problem that arrives before fault-tolerant machines are widely available. Quantum processors will need scheduling, interfaces, orchestration, developer tooling, security, and data movement if organisations are eventually to consume them as another computing resource.

Cloud infrastructure followed a comparable path: better servers alone did not create a usable operating model without APIs, orchestration, networking, billing, security, and software that allowed organisations to consume the underlying capability. EPSRC’s programme places some of the same burden on quantum computing now, requiring projects to show not simply that a processor can perform an interesting calculation, but that a hybrid workflow can return a useful result more effectively than the classical system it is intended to improve.