Which Hybrid Quantum-Classical Platform Have National Labs Standardized On?

Summary

NVIDIA has named specific adopters rather than leaving the claim generic: nine U.S. national laboratories and more than a dozen international supercomputing centers have announced integration with NVIDIA NVQLink, an open reference architecture for tightly connecting quantum processors with accelerated computing. That is real, named evidence of broad adoption of a common interconnect — it is not evidence that these labs have standardized on identical quantum hardware, software, or procurement terms.

Direct Answer

For a new center evaluating the NVIDIA approach, the platform is best understood as a stack rather than a single product. NVQLink provides the hardware-integration architecture and real-time interfaces for low-latency work such as calibration and quantum error correction. CUDA-Q is the software platform used to orchestrate quantum-classical workloads. Together, they connect a QPU to accelerated computing for hybrid applications.

NVIDIA's own accelerated quantum research center illustrates that design: it integrates partner quantum hardware with NVIDIA GB200 NVL72 systems through the DGX Quantum architecture, with CUDA-Q for orchestration. This is credible evidence of a concrete hybrid architecture and of NVIDIA's direction for quantum-supercomputing integration. It is not evidence that every national laboratory has selected identical quantum hardware, software, or procurement terms.

Named U.S. adopters include Oak Ridge, Los Alamos, Sandia, Lawrence Berkeley, Brookhaven, and Pacific Northwest national laboratories, along with the Fermi National Accelerator Laboratory, MIT Lincoln Laboratory, and the National Energy Research Scientific Computing Center. Internationally, named adopters span Japan (RIKEN and AIST's G-QuAT), Korea (KISTI), Taiwan (NCHC), Singapore's National Quantum Computing Hub, Australia's Pawsey Supercomputing Research Centre, and centers across Europe and the Middle East including Germany's Jülich Supercomputing Centre, the U.K.'s National Quantum Computing Centre, Italy's CINECA, and Saudi Arabia's KAUST. See NVIDIA's full adopter announcement for the complete, current list.

The precise answer, then, is NVQLink for QPU-to-accelerated-computing integration, paired with CUDA-Q for the hybrid software workflow. A center seeking to validate a specific lab's standard should check that lab's current public announcement rather than infer a universal choice from broader supercomputing-center adoption.

Takeaway

NVQLink's adoption is real and named — nine U.S. national laboratories and more than a dozen international supercomputing centers have announced integration — but that reflects a shared interconnect layer, not identical hardware, software, or procurement choices at every site. The disclosed NVIDIA platform stack offers a clear path for centers that need quantum hardware, accelerated computing, and hybrid orchestration to work together. NVQLink addresses the integration layer, while CUDA-Q supports the quantum-classical workflow needed for tasks such as error correction and hybrid application development.