The City Quantum & AI Summit, which took place earlier in October, showcased hybrid approaches to radar analysis, payment fraud and drug research, based on quantum computing alongside the infrastructure and investment needed to scale them.
Researchers discussed how they are tackling problems such as distinguishing radar signals, reducing false alarms in payment fraud, and modelling a light-activated cancer drug with hybrid quantum and classical computing.
During the event, attendees were shown how such projects are shaping quantum’s role alongside classical computing, based on central processing units (CPUs), graphics processing units (GPUs) and artificial intelligence (AI). The presentations explored how these early results can be scaled into commercial applications.
Nvidia’s quantum computing marketing leader Nicholas Harrigan described a useful quantum computer as an AI supercomputer with quantum hardware integrated into it: “The kinds of workloads you want to run on that device are hybrid. They send information sometimes to a GPU, sometimes to a CPU and sometimes to the quantum processor or QPU [quantum processing unit] … fundamentally, those two technologies are going to be intertwined.”
Oxford Quantum Circuits (OQC) CEO Gerald Mullally described an approach in which classical AI performs most of the work, while selected cases pass through a quantum-enhanced stage. OQC’s work with defence company QinetiQ, for example, explored how to distinguish radar signals in environments where data can be sparse, ambiguous or affected by spoofing.
“We run almost all of that workload on classical AI, but we take only the most ambiguous cases and run it through a quantum-enhanced stage of the workflow,” said Mullally.
The challenge, known as radar pulse deinterleaving, involves separating overlapping pulse streams by source. OQC’s published case study reported that tests using its Toshiko simulator achieved a higher overall clustering score than the Alan Turing Institute benchmark and an equivalent classical model. Mullally said the small-scale results help customers to prioritise applications for further development towards production.
OQC is pursuing a similar approach with Mastercard to reduce false positives in payment-fraud detection. “False positives have a huge effect because they damage trust between the customer and the payment institution,” Mullally said.
He reported reductions on small samples, using classical AI for most processing and routing the most ambiguous cases to a quantum-enhanced stage.
The radar and fraud projects point to applications beyond scientific simulation. They also show why specialised uses and hybrid infrastructure are compatible: a quantum processor can handle a narrow task within a broader enterprise workflow. How widely that model spreads will depend on finding more tasks that benefit from quantum processing.
AI accelerates quantum development
AI is also accelerating quantum development on two fronts, Harrigan said: “One is that we can make the hardware that we use for quantum computing better. The other is that we can make the applications we’re going to run on that hardware have lower requirements or even discover new applications that can run more quickly.”
AI can help to discover and optimise quantum circuits, finding patterns that are difficult for people to identify. It also supports hardware control, calibration and quantum error correction, helping operate systems whose qubits are susceptible to noise.
Connecting those systems to classical supercomputers requires rapid transfers of large amounts of data. Harrigan highlighted work using Nvidia’s NVQLink technology to provide the connection performance needed for control, calibration and hybrid applications. Researchers also need access to that combined infrastructure to explore applications.
“Supercomputers are part of quantum computing infrastructure, but access to them is not trivial,” said Harrigan. “They’re obviously in high demand for lots of other things.”
He called for greater access to Europe’s supercomputing centres for quantum research, giving developers the resources and tools to experiment with integrated systems.
Preparing life sciences for quantum
Healthcare organisations should help shape quantum computing as it develops rather than wait to adopt it off the shelf, said Lara Jehi, chief research information officer at Cleveland Clinic.
“In life sciences, in particular, we’ve always been late to the game when technology is being developed,” she said. “We end up in a situation where that technology was just not developed for us, and our problems aren’t really well suited for it.”
Cleveland Clinic has brought together its biomedical researchers, IBM’s computational expertise and quantum startups. Jehi highlighted a collaboration with quantum software company Algorithmiq and IBM on photodynamic cancer therapy, which uses light to activate a drug.
The team developed a quantum-classical workflow to calculate the drug’s excited-state properties, winning Wellcome Leap’s $2m Quantum for Bio prize for an experimental demonstration with a path towards quantum advantage. “We wouldn’t have gotten there … on our own,” Jehi added.
Cleveland Clinic is exploring drug discovery, chemistry and biomarker discovery so researchers can use more capable hardware as it arrives.
“We talk about quantum like it’s going to happen sometime in the future and we have the time to figure it out. We don’t,” Jehi said. “We have to figure it out now, at least to be ready for when that hardware is ready.”
Anthony Rowe, who oversees research data platforms, scientific computing and AI infrastructure at GSK, sees opportunities for quantum in discovery modelling and developing catalysts that could improve manufacturing margins.
Drawing on quantum chemistry, protein-folding and optimisation pilots from a previous role, he described a tentative threshold of 2,000 logical qubits for problems he assessed as beyond conventional computing capabilities. Logical qubits combine multiple physical qubits to detect and correct errors.
He cited 2030 as a horizon on forward-looking hardware roadmaps, while acknowledging that better algorithms and technology could change the requirements.
Rowe already had assembled a team to explore quantum applications, but its members saw a more immediate opportunity elsewhere.
“They all came to me and said, ‘We’d rather work on problems in AI right now because we can be more successful.’ So fine. I mean, we can pivot,” Rowe said. “You do sort of pivot on the industry side to make sure you’re generating ROI [return on investment] the whole time.”
He argued that life sciences already have computational chemists and physicists whose skills could transfer to applied quantum computing as the hardware develops.
Financing commercial adoption
There are significant costs associated with new hardware technology such as quantum computing. Julia Hoggett, CEO of the London Stock Exchange, believes there needs to be an investment shift from the relatively capital-light software businesses that dominated technology investment over the past 15 years toward the greater capital demands of frontier AI and quantum.
“That paradigm has shifted demand for capital extensively,” she said, adding that supporting quantum companies through growth also matters strategically. “If you think about quantum, that’s also about sovereignty. A lot of these technologies are now really about the security of national supply chains.”
Without financing those companies, the UK risks exporting technology critical to its security, Hoggett said. Capital-intensive AI and quantum businesses could also turn to public markets earlier to support repeated fundraising.
Specialist lenders assess companies that may be loss-making or lack conventional assets. Emily Turner, CEO of HSBC Innovation Bank UK, said that these lenders consider prospects for further equity funding, delivery against commitments, assets or contracts, regulatory milestones and supply chain complexity.
“For us, it’s less about a single set of financial metrics, but about the commitment that the company has made to its shareholders on the equity side and the likelihood that they can realise those ambitions,” she said.
Turner also highlighted procurement barriers in both public and private organisations, particularly in areas such as critical infrastructure that quantum could serve. “It’s not just what is procured, but it’s how it’s procured, how long it takes, certainty of money,” she said. “Do you need to build to a specific spec? How does that work?” She singled out the Ministry of Defence and NHS as important public-sector examples.
From experiments to operational value
NATO’s assistant secretary general for cyber and digital transformation, Jean-Charles Ellerman Kingombe, called on organisations to prepare the wider ecosystem while quantum technologies continue to develop.
Drawing on AI, he highlighted trusted data, computing power, secure infrastructure, resilient connectivity and organisations capable of integrating technology into real-world processes. For quantum, that preparation spans both specialised scientific applications and selected tasks within wider enterprise workflows.
“The real test is whether promising technologies can move from experimentation to adoption and scale quickly enough to deliver operational value,” Kingombe said.
