In February 2026, IBM and Japan's RIKEN demonstrated what the future of quantum computing may actually look like: not a quantum computer replacing a supercomputer, but quantum and classical machines working continuously together.

The researchers connected RIKEN's IBM Quantum Heron processor with Fugaku, one of the world's most powerful classical supercomputers, and operated them in a closed-loop workflow. The two systems repeatedly exchanged information while calculating the electronic structure of complex iron-sulfur molecules.

Fugaku contains 158,976 computing nodes, each built around a 48-core processor. Rather than running the quantum and classical calculations independently, IBM and RIKEN developed an orchestration system that allowed the machines to continuously feed results back to one another while minimizing idle time.
The workflow used sample-based quantum diagonalization (SQD). The quantum processor helps sample the enormous space of possible electronic configurations in a molecule, while the classical supercomputer performs the computationally intensive calculations needed to refine the resulting energy estimates.
IBM described the result as the largest and most accurate chemistry calculation yet conducted using a quantum computer. The calculations exceeded what could practically be achieved through exact classical diagonalization and produced accuracy comparable with sophisticated classical approximation methods.
Importantly, this is not yet proof that the quantum computer outperformed every classical alternative. Approximate classical methods remain extremely competitive.

The greater significance is infrastructure. Many practical quantum algorithms are expected to require CPUs and GPUs to perform substantial portions of the computation. A useful quantum processor may therefore become a specialized accelerator inside a much larger high-performance-computing environment—similar to how GPUs complement CPUs today.

The IBM-RIKEN experiment provides an early blueprint for that architecture.Instead of asking when quantum computers will replace conventional supercomputers, the more realistic question may be when quantum processing units become standard components inside future supercomputers.