Yonsei to Run IBM Nighthawk Against Leigh Syndrome

Leigh syndrome is a rare genetic disease that attacks the brain and nervous system, and it is merciless. Most children diagnosed with it do not live past their third birthday. Working out the biological mechanism behind it means modelling molecular interactions so fiendishly complex that running Japan’s Fugaku, one of the fastest supercomputers on Earth, without a single break, would take an estimated 48 years to finish the sums.

That is the problem a team at Yonsei University in South Korea is pointing a quantum computer at. On 12 August the Seoul Economic Daily reported that the Yonsei Quantum Initiative will replace its current IBM machine with a next-generation IBM Nighthawk processor in November, making the Songdo campus in Incheon the world’s second Nighthawk site after IBM’s own facility in Miami.

Table of Contents

What Nighthawk actually changes

The upgrade is about wiring, not raw qubit count. Initiative director Jung Jae-ho explained the difference plainly. On the older Eagle chip, each qubit connected to roughly 2.5 of its neighbours on average, so getting two distant qubits to talk required a chain of “swap” operations, and errors piled up at every step. Nighthawk uses a lattice layout in which each qubit links directly to four neighbours, cutting out a lot of those swaps.

The payoff, according to Jung, is about 40% more computation than Eagle at the same error level. That is a meaningful jump, though nobody at Yonsei is pretending the machine is error-corrected. IBM itself reckons the real inflection point arrives around 2029, when large-scale error-corrected systems are meant to exist. Until then the game is squeezing useful work out of noisy hardware.

Quantum and classical, working the same problem

The Leigh syndrome research is a hybrid job. The plan is to let the quantum computer narrow down the most promising molecular candidates first, then hand those off to Fugaku, RIKEN’s supercomputer in Japan, for the heavy classical number-crunching. Do it that way, Jung said, and a calculation that would take Fugaku 48 years on its own could drop to a matter of days. Yonsei planned to review the first quantum results in August and discuss the next round with RIKEN.

This is the honest version of what quantum computing can do today. It works as a clever scout for the classical machine, telling it where to dig so it does not waste decades wandering down dead ends.

A bridge for companies that are lost

Alongside the hardware, Yonsei is launching a platform called Q-Bridge, due to be finished in early November to line up with the Nighthawk install. The pitch is aimed at firms that suspect quantum computing might help them but have no clue which algorithm to reach for or how to split a problem between quantum and classical systems. As Jung put it, most companies have no quantum literacy, so the idea is to lower the barrier and get them to bring their real-world problems to the table first.

Useful cases from Q-Bridge would then be stored in a resource called the Q-Library and reapplied to similar problems later. There is also a tie-up with the University of Cambridge, including plans for each side to open a branch on the other’s campus, linking Yonsei’s drug-research work with Cambridge’s Milner Therapeutics Institute.

None of this makes South Korea a quantum superpower overnight, and a working, error-corrected machine remains years off. But it is a real, deployed processor aimed at a real disease, which is a good deal more grounded than most of the world-first noise this field kicks up. A child with Leigh syndrome does not care which qubit topology finds the answer. Speed is the entire point.

Did you know: IBM names its quantum processors after birds, Eagle, Heron, Condor and now Nighthawk, a scheme it has run for years, going back to the 27-qubit Falcon chip.

Sources

The free stack. One email a week: the AI tools and moves that actually matter, hype filtered out. Subscribe free →