February 18, 2026 -- RIKEN and IBM have achieved a new milestone in quantum-centric supercomputing (QCSC). A joint research team used the entirety of one of the world’s leading high-performance computing systems—the pre-exascale supercomputer Fugaku—in concert with RIKEN's on-premises IBM Quantum Heron processor to calculate the electronic structure of a pair of complex iron-sulfur molecules. The result was the largest and most accurate chemistry experiment ever performed on a quantum computer, which IBM Director of Research Jay Gambetta presented 29 Jan at Supercomputing Asia 2026.

This work, described in an October 2025 paper on the arXiv, is the first demonstration of quantum and HPC working together this closely at this scale. Orchestrating them required the development of a new closed-loop workflow, where the supercomputer and quantum computer process data and feed results back and forth to one another. This represents a major milestone for QCSC, and a significant technical achievement in making efficient use of both the quantum computer and the immense power of Fugaku with minimal execution time.

“This is a very exciting development for hybrid computing,” said Mitsuhisa Sato, Division Director of the Quantum-HPC Hybrid Platform Division at RIKEN Center for Computational Science.

Together, the supercomputer and quantum computer ran a workflow based in sample-based quantum diagonalization (SQD). A recent IBM Research blog predicts that different kinds of compute resources—including CPUs, GPUs and QPUs—will increasingly work together to enable and accelerate hybrid quantum-classical algorithms like SQD.

Fugaku is an enormous supercomputer, made up of 158,976 chips, each with 48 cores. It reigned as the fastest computer in the world from 2020 to 2021. QCSC will require architectures that integrate quantum computers with this scale of classical hardware, and workflows to coordinate them. Another recent RIKEN-IBM paper on the arXiv shows how adding GPUs to the workflow could dramatically accelerate this sort of execution. The quantum and HPC communities are coming together to integrate these technologies to realize quantum advantage today.

Until recently it’s been typical for classical and quantum HPC resources to be used sequentially: the quantum processing unit (QPU) does some work, then the results get shuttled over to the classical computer which does some work and sends the results back. Fugaku and RIKEN’s on-premises Heron worked together in an orchestration more closely resembling a practical implementation of quantum in an HPC environment.

How orchestration drives advantage

Orchestrating classical and quantum compute resources in a single workflow is a very complicated problem, particularly when the interdependency and complexity of the quantum and classical parts of the workflow are substantial. Implementing SQD across Heron and a supercomputer the size of Fugaku isn’t as simple as plugging in the algorithm and letting the computers go to work. Quantum-centric supercomputing (QCSC) requires defined workflows and tight orchestration of quantum and classical resources. The researchers had to figure out how to make a closed loop work at this scale.

Efficient orchestration matters because quantum and classical resources are precious, said Hiroshi Horii, IBM Senior Manager of QCSC and Head of IBM Quantum Japan. If your workflow is inefficient, and either resource, classical or quantum, is left idle at any point during the computation, you are wasting precious runtime that other researchers could use to tackle other important computations. Fugaku is a billion-dollar machine, and every second of its uptime is valuable. It’s crucial that those seconds aren’t wasted sitting around waiting for Heron to finish up a step in the calculation.

To this end, the researchers developed a new task assignment system. As Fugaku and Heron performed iterative calculations, passing information back and forth in a closed loop, this system assigned tasks to make sure both computers were working as much as possible, minimizing the time-to-solution. While this workflow was designed specifically for Fugaku, they found it could also be applied to a range of cloud HPC environments—showing that quantum computers can interact efficiently with classical HPC infrastructures. Researchers now have a better understanding of how to run quantum computers alongside classical HPC at this scale.

Modeling iron-sulfur molecules with a powerful quantum-centric algorithm

Using Fugaku and Heron together, the researchers calculated the electronic structure of a pair of iron-sulfur molecules. Calculating their electronic structure means figuring out how their electrons are distributed and how they behave. Understanding the behavior of a molecule’s electrons tells you a great deal about how it will interact and react with its environment.

As part of the effort to realize the full potential of QCSC, the community is developing a new class of algorithms. These algorithms break problems into parts best handled by quantum resources and parts best handled by classical resources. The quantum computer acts like the lifting pin in a lockpicking set, unlatching the most complex part of the problem. The classical computer turns the handle and opens the door.

SQD is part of the new class of algorithms for QCSC. It addresses one of the most fundamental challenges of electronic structure calculations: the total space of possible arrangements for a molecule’s electrons is enormous, and that space only grows as molecules get more complex. In SQD, the quantum computer samples that space, identifying key areas for the classical computer to focus on. The classical computer uses that information to drive toward a solution.

Looking ahead to quantum-centric supercomputing

With all of Fugaku working in a closed loop with Heron, the results were remarkably accurate—much more accurate than several previous attempts to solve the same problem with quantum computers. This work is beyond the reach of exact classical methods; and comparable to some of the most advanced classical approximation methods.

The next step, the researchers said, is integrating GPUs as accelerators of quantum-classical workflows. As this work scales, said Tomonori Shirakawa, senior research scientist at RIKEN, quantum advantage is on the horizon.

“Could we see it at RIKEN this year?” He replied, “We need to see that effort, but I’m very optimistic.”