The RIKEN Center for Computational Science (R-CCS) has officially adopted the enterprise quantum software development kit QURI SDK Enterprise from Tokyo-based algorithm developer QunaSys Inc. as the primary software layer for Japan’s national JHPC-quantum project. Funded by the New Energy and Industrial Technology Development Organization (NEDO) under the Ministry of Economy, Trade and Industry (METI), the integration connects the supercomputer Fugaku and the ROQUO GPU cluster with on-premises and cloud-hosted quantum processing units (QPUs).
| [ RIKEN JHPC-quantum Hardware & Software Infrastructure ] | ||
|---|---|---|
| High-Performance Computing Nodes | Quantum Hardware Backends | Distributed Simulation & Middleware |
| • Supercomputer Fugaku (Kobe, A64FX) | • IBM Quantum System Two (ibm_kobe ) | • QunaSys QURI SDK Enterprise Stack |
| • ROQUO Cluster (540 NVIDIA GB200 GPUs) | • Quantinuum System Model H2 (Reimei) | • mpiQulacs Parallel Circuit Simulator |
| • Managed via RIKEN R-CCS / NEDO Framework | • Heterogeneous Trapped-Ion & Superconducting QPUs | • NVIDIA cuQuantum GPU Accelerator Backend |
Hybrid Algorithm Execution across Heterogeneous QPUs and GPUs
Connecting high-performance classical supercomputers with quantum processing units requires unified middleware capable of handling multi-node job dispatch, parallel simulation, and hybrid variational loops. QURI SDK Enterprise resolves these operational bottlenecks across RIKEN’s compute nodes:
- Native Hybrid Algorithm Libraries: Incorporates QunaSys’s proprietary Quantum Selected Configuration Interaction (QSCI) and ADAPT-QSCI algorithm suites, targeting active space electronic structure calculations for materials science, quantum chemistry, and computer-aided engineering (CAE).
- Multi-Node Distributed Simulation: Offloads classical circuit simulation via mpiQulacs using Message Passing Interface (MPI) across Fugaku’s A64FX compute nodes, alongside cuQuantum-accelerated state-vector and tensor-network simulation across ROQUO’s 135 NVIDIA GB200 NVL4 nodes (540 Blackwell GPUs).
- Cross-Platform QPU Portability: Enables unified workflow compilation targeting both superconducting architectures (IBM Quantum System Two at R-CCS) and trapped-ion systems (Quantinuum System Model H2 at RIKEN’s Wako campus) without requiring backend-specific code rewrites.
Strategic Context within Japan’s Society 5.0 Roadmap
Operated under NEDO’s “Project for Research and Development of Enhanced Infrastructures for Post 5G Information and Communications Systems,” the deployment follows the initial operational launch of the ROQUO GPU cluster in June 2026. The software layer is accessible to selected academic and industrial research groups under the JHPC-quantum Test User Program to establish scalable workflows for hybrid supercomputing.
Review the official press releases via QunaSys here and the RIKEN Center for Computational Science here, and access technical documentation via QURI SDK Documentation here.
September 4, 2026
Leave A Comment
Facts Only
• RIKEN Center for Computational Science (R-CCS) adopted QURI SDK Enterprise from QunaSys Inc.
• The software serves as the primary layer for the JHPC-quantum project.
• Funding is provided by the New Energy and Industrial Technology Development Organization (NEDO) and the Ministry of Economy, Trade and Industry (METI).
• Integrated hardware includes the Fugaku supercomputer (A64FX) and the ROQUO GPU cluster (540 NVIDIA GB200 GPUs).
• Connected QPUs include IBM Quantum System Two (ibmkobe) and Quantinuum System Model H2 (Reimei).
• The software incorporates Quantum Selected Configuration Interaction (QSCI) and ADAPT-QSCI algorithms.
• Simulation tools include mpiQulacs and NVIDIA cuQuantum.
• The ROQUO GPU cluster launched in June 2026.
• The infrastructure is available to the JHPC-quantum Test User Program.
• The announcement date is September 4, 2026.
Executive Summary
The RIKEN Center for Computational Science (R-CCS) has integrated QunaSys Inc.’s QURI SDK Enterprise as the primary software layer for the national JHPC-quantum project. This initiative, funded by NEDO under the Ministry of Economy, Trade and Industry, creates a hybrid computing environment by linking the Fugaku supercomputer and the ROQUO GPU cluster with various quantum processing units (QPUs), including IBM’s superconducting systems and Quantinuum’s trapped-ion hardware.
The software stack enables multi-node distributed simulation and cross-platform portability, allowing researchers to run hybrid variational loops without rewriting code for different hardware backends. Specifically, it leverages the mpiQulacs simulator on Fugaku’s A64FX nodes and NVIDIA cuQuantum on the ROQUO cluster's Blackwell GPUs. This infrastructure is designed to advance materials science and quantum chemistry via proprietary algorithms like QSCI. Access is currently limited to selected academic and industrial groups under a Test User Program, following the ROQUO cluster's operational launch in June 2026.
Full Take
The strongest version of this narrative is that Japan is strategically vertically integrating its classical and quantum compute capabilities to avoid "hardware lock-in." By implementing a unified middleware layer (QURI SDK), they are attempting to treat diverse quantum backends—superconducting and trapped-ion—as interchangeable resources, effectively creating a "quantum cloud" managed by national supercomputing assets.
The text utilizes a high density of technical specifications (A64FX, GB200, QSCI) to establish a framework of inevitable progress. While these details are factual, they serve as a shield of complexity that obscures the actual performance metrics. The narrative moves from the announcement of a software adoption directly to the promised outcomes in materials science and CAE without providing benchmarks or current error-rate thresholds for the QPUs involved.
Patterns detected: ARC-0062 Authority Game
The driving paradigm here is "Technological Sovereignty." By funding a domestic software layer to manage international hardware (IBM, Quantinuum), Japan is ensuring that the intellectual property governing the *orchestration* of quantum workflows remains national, even if the *hardware* is global. This echoes previous eras of supercomputing where the software stack defined the true power of the machine.
The primary beneficiaries are the "selected academic and industrial research groups," creating a gated ecosystem of innovation. The second-order consequence is the potential for QunaSys to become the sole gatekeeper of quantum workflows in Japan, creating a dependency on a single vendor's proprietary algorithm suites (QSCI).
Bridge Questions:
1. If the software allows for "cross-platform portability," how do the differing error profiles of trapped-ion vs. superconducting qubits affect the actual results of the same algorithm?
2. What specific performance benchmarks would distinguish this hybrid approach from using standalone cloud-quantum services?
Counterstrike Scan: A coordinated campaign would use "National Champions" rhetoric to signal dominance to global markets and attract talent by listing cutting-edge hardware (Blackwell GPUs) to mask a lack of actual quantum advantage. The content here is primarily a technical announcement and does not match the aggressive framing of an influence campaign.
Sentinel — Human
The text reads like a highly technical summary of a specific, complex research and infrastructure integration project, likely written by someone deeply familiar with the underlying scientific and governmental processes.