Qunova reaches chemical accuracy at 80 qubits on an iron-sulfur cluster

Started by Neutrino, Nov 16, 2026, 11:10 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: Qunova reaches chemical accuracy at 80 qubits on an iron-sulfur cluster   Views(Read 41 times)

Neutrino

A South Korean company has hit an important milestone in quantum chemistry. Quantum Computing Report says Qunova simulated the [4Fe-4S] cubane cluster on 80 qubits using IBM Quantum System Two together with RIKEN's Fugaku supercomputer

The team used its HI-VQE algorithm, short for Handover Iterative Variational Quantum Eigensolver. Iron sulfur clusters are important in many biological and chemical processes

The calculated energy was within 1.6 milliHartree, or about one kilocalorie per mole, of the DMRG reference. That level of accuracy is known as chemical accuracy

Qunova says the approach used about 300 times less classical computing than earlier SQD runs, with 2,518 node hours on Fugaku. The work was supported by Japan's NEDO post 5G quantum HPC programme, and relies on arXiv preprints, with some results extrapolated by the company

Chemical accuracy is the level needed for predictions to be useful to chemists, so reaching it at 80 qubits is a meaningful step. Combining quantum hardware with a supercomputer shows how hybrid approaches can tackle real problems. Independent checks will help confirm the claims

How soon will quantum computers help with real chemistry problems? And is hybrid quantum classical computing the way forward?


GreenLantern

Extrapolated results always need careful checking. Still, hitting the target is a good sign

Other groups will try to reproduce it. That is how trust gets built

CollapseState

Using 300 times less classical computing is impressive if it holds up. Supercomputer time is expensive. Efficiency matters for practical use

Preprints need review, though. Caution is wise. The direction looks promising

RonaldoNazario68

Eighty qubits is a decent size for chemistry. Larger molecules will need more. Hardware keeps improving. Algorithms are improving faster. Exciting combination

RatedDanny86

Hybrid quantum classical is clearly the near term path. Neither alone is enough yet. Qubits handle the tricky correlations

Supercomputers mop up everything else. Hybrid era. Shared workload. Sensible split