How Rolls-Royce is testing quantum computing for jet engine design
Rolls-Royce is moving from emulation to quantum hardware as it explores whether fault-tolerant machines could transform jet-engine design
Rolls-Royce has spent six years writing algorithms for computers that do not yet exist. Its four-person quantum team believes fault-tolerant machines could eventually help engineers simulate more than a trillion variables in jet-engine air-flow models, but it’s now moving to real quantum hardware.
How quantum computing could improve jet-engine design
Rolls-Royce began its quantum computing programme in 2020, hoping, in particular, to bring the nascent technology to bear on the design of gas turbines, where the flow of gas is modelled using computational fluid dynamics (CFD) simulations.
The aerospace giant’s largest research CFD models, which run on high-performance computers (HPC), already solve equations involving several billion variables, but it wants eventually to go beyond a trillion – which is where quantum computing could help.
“Aero engines are already very efficient, so a lot of it is new design architectures,” said Rolls-Royce fellow of computational science Leigh Lapworth.
He pointed to UltraFan, Rolls-Royce’s next-generation engine technology programme. “What we're doing is guided towards that, where we want to do a lot more design by analysis.”
Rolls-Royce is betting on fault tolerance
The quantum team, led by Lapworth, decided from the outset to target fault-tolerant quantum computing. Instead of building for the already available NISQ (noisy intermediate-scale quantum) devices, it chose to design algorithms for fully error-corrected quantum computers once they become a reality. In 2020, the expectation was that error-corrected devices were about 10 years away. “We were very much an outlier making that decision,” Lapworth said.
But vendor timelines have since become more ambitious. If the roadmaps of vendors including QuEra, PsiQuantum and Xanadu are correct, early error-corrected machines could emerge from 2028, with some vendors like IBM and Quantinuum promising large-scale fault-tolerant systems around 2029-2030, although definitions of such terms differ and these are estimates rather than guarantees.
Lapworth’s core team is tiny, only four people, but three Innovate UK-funded projects have allowed it to work with roughly 30 specialists from organisations including Riverlane, Universal Quantum, Xanadu and Classiq. Lapworth says this partnership model gives the company considerable leverage. The team also publishes papers and makes test cases allowing outside researchers to get involved.
“Although we're small, we're doing a lot with quantum computing experts and scientists,” he said.
From quantum emulation to real hardware
Until now, the team has developed and tested algorithms using emulators rather than quantum hardware. The programme is now moving into a second phase, in which Rolls-Royce plans to evaluate algorithms on Quantinuum hardware, working alongside Riverlane (quantum software, error correction) and the University of Edinburgh (HPC, quantum-classical workflow integration).
“Everything we've done so far has been in emulation, because... there hasn't been a computer to run on, so we're now transitioning to what can we actually do,” said Lapworth. He stressed that the work remains experimental rather than operational, placing it at around Technology Readiness Level 3 – experimental proof of concept - on NASA’s nine-level scale.
The programme is following what Lapworth calls a three-times-five strategy, or three phases each of five years. “We are now in this transition from phase one to phase two. Phase two is the hardware evaluation, so we're going to work with Quantinuum on that. Phase three will focus on industrialisation.”
Importantly there has been significant progress in the underlying technology. Lapworth pointed to Google’s 2024 Willow paper, which showed below-threshold error correction, meaning as it increased the size of its surface code, the logical error rate fell.
The importance of public funding
Government support for quantum experimentation is "absolutely crucial", Lapworth said. Rolls-Royce has completed three Innovate UK-backed projects and continues to seek support through programmes including Commercialising Quantum. UK funding can also support international partnerships, including bilateral arrangements with Canada, Germany and Denmark.
Backing by the UK's national innovation agency, Innovate UK, has allowed the four-person internal team to collaborate with quantum expertise in a way that would otherwise be out of reach. Without ring-fenced funding his small team, with its distant commercial goals, would be competing with other projects within Rolls-Royce offering more immediate returns.
“We’d be towards the bottom of the queue,” Lapworth said. “Innovate UK has been hugely influential and done a lot of good for the country.”
Asked how he selects the companies he works with, Lapworth said relationships have generally developed organically through conversations with like-minded researchers.
The field has been sharply narrowed by Rolls-Royce’s focus on fault tolerance. That made Riverlane an obvious early partner, because of its work on error correction, and subsequent partnerships with Xanadu and Classiq grew from that relationship.
“A lot of it was personal chemistry,” said Lapworth, who credited Riverlane CEO Steve Brierley and Universal Quantum CEO Seb Weidt as influential voices.
Staying platform-agnostic as qubit technologies compete
Rolls-Royce isn't trying to pick a winning qubit modality just yet. Rather, it is developing algorithms that can exploit whichever architectures deliver ultimately useful fault-tolerant computing.
Lapworth said he’s encouraged by progress in several fields including trapped-ion systems (like Quantinuum) which offer very high fidelities, and the highly scalable neutral-atom approach. Ultimately, though, he thinks error correction capabilities will outweigh most considerations.
“The speed of a machine won't necessarily be the speed of the qubits, it will be the speed of the error correction.”
AI could become an important accelerant too, helping with areas such as circuit compilation, error-correcting codes and hardware calibration.
Lapworth expects the biggest challenge will be industrialisation: standardising processes and validating them to aerospace specifications, so “the results we get doesn't depend on the engineer that actually does that piece of work”.
Therefore, he expects quantum adoption to be gradual, used in combination with HPC, and potentially stretching well into the next decade rather than arriving at a single "quantum advantage" milestone any time soon.