IBM released the industry’s first published quantum-centric supercomputing reference architecture, outlining a scalable blueprint to combine quantum processors (QPUs) alongside traditional GPU and CPU clusters. This design supports deployment across on-premises data centers, cloud platforms and research facilities, aiming to tackle scientific challenges beyond the reach of classical systems alone.

The architecture integrates high-speed networking and shared storage with coordinated orchestration tools, enabling seamless data exchange between quantum and classical components. Open software frameworks, including Qiskit, allow developers to schedule and manage hybrid workflows through familiar programming environments and toolchains.

Researchers have used the quantum-centric system to simulate a half-Möbius molecule, a 303-atom tryptophan-cage mini-protein and iron-sulfur clusters, delivering results that outperform classical-only approaches. Collaborations with institutions such as the University of Manchester, ETH Zurich and RIKEN highlight the architecture’s capability to accelerate molecular and materials research.

IBM plans to evolve the reference design through its global ecosystem of clients and partners, enhancing orchestration and algorithm support. A collaboration with Rensselaer Polytechnic Institute will refine workflow scheduling across quantum and high-performance computing resources, positioning the architecture to drive applications in chemistry, materials science and optimization at scale.