Rebeca Moen
Nov 18, 2025 03:32
NVIDIA introduces NVQLink, an open platform architecture that integrates accelerated computing with quantum processors, enhancing the capabilities of quantum computing systems.
NVIDIA has unveiled NVQLink, a pioneering architecture that integrates accelerated computing with quantum processors, according to NVIDIA’s blog. This development marks a significant step forward in the evolution of quantum computing, as it aims to meet the demanding computational requirements of quantum processing units (QPUs).
Revolutionizing Quantum Computing
NVQLink is designed to seamlessly integrate conventional supercomputing hosts with quantum system controllers (QSCs), enabling GPU superchips to support the online workloads of QPUs. This open platform architecture is compatible with existing control systems, including superconducting, trapped-ion, photonic, or spin-based systems. By making the supercomputing node a native part of the QPU environment, NVQLink accelerates the computational capabilities of quantum hardware.
System and Software Architecture
The NVQLink architecture introduces a machine model known as the Logical QPU, which encompasses physical qubits, control and readout electronics, and compute resources for online workloads like quantum error-correction (QEC) decoding. This hybrid system combines quantum coherent control with state-of-the-art supercomputing, connected by a low-latency, scalable real-time interconnect.
Developers can leverage the CUDA-Q platform to program the real-time host in C++ or Python, while a third-party QSC manages qubits’ control through an array of FPGAs or RFSoCs. The real-time interconnect, implemented using RDMA over Ethernet, facilitates rapid data exchange with latency under four microseconds.
Enhancing Real-Time Quantum Applications
Through NVQLink, the QPU transforms from a peripheral device into a first-class peer within a supercomputer, enabling a hybrid accelerated quantum supercomputer. Developers benefit from a unified heterogeneous programming model, allowing them to write programs that express quantum kernels and real-time callbacks.
NVIDIA’s collaboration with Quantinuum demonstrates NVQLink’s potential, with their Helios QPU utilizing NVIDIA’s GH200 Grace Hopper as a real-time host for quantum error correction. The integration allows for real-time decoding and error correction using NVIDIA’s nv-qldpc-decoder, showcasing a significant reduction in error rates.
Real-World Applications and Future Prospects
The integration of NVQLink into quantum computing systems offers a scalable solution for quantum and traditional computation to coexist, benefiting from advancements in RDMA technology. As large AI and supercomputing deployments continue to drive innovation, NVQLink is well-positioned to leverage these developments for quantum systems.
NVIDIA’s NVQLink architecture represents a transformative approach to quantum computing, providing a foundation for faster experimentation and the development of scalable quantum systems. This open platform is set to accelerate the roadmap for QPU builders, researchers, and operators alike.
Image source: Shutterstock
Source: https://blockchain.news/news/nvidia-nvqlink-bridges-accelerated-and-quantum-computing
