Insider Brief
- Quantum Machines announced its integration with NVIDIA’s NVQLink platform, advancing real-time orchestration between quantum and classical computing systems.
- The integration extends the DGX Quantum architecture, enabling microsecond-latency data exchange between quantum processors, GPUs, CPUs, and control hardware for error-corrected quantum computing.
- NVQLink allows seamless hybrid execution across NVIDIA’s CUDA-Q platform, with Quantum Machines’ OPX hardware enabling deterministic quantum-classical feedback on production systems.
PRESS RELEASE — Quantum Machines (QM), the leading provider of quantum control solutions, today announced its integration with NVIDIA NVQLink, the new open platform for real-time orchestration between quantum and classical computing resources. This marks a major step that extends QM’s first-of-its-kind, field-proven, µs-latency quantum-classical integration solution.
Building on the foundation of NVIDIA DGX Quantum – the first system to connect a quantum controller directly with the NVIDIA accelerated computing stack – QM’s platform will support the new NVQLink open architecture, providing seamless interoperability between quantum processors (QPUs), control hardware, CPUs, and GPUs. The result is real-time data exchange and control at microsecond latency, enabling the demanding workloads required for logical qubits and large-scale quantum error correction.
“Our team worked closely with NVIDIA to shape the NVQLink specifications, based on all our insights from the last several years of working on quantum-GPU integration,” said Itamar Sivan, CEO and Co-founder of Quantum Machines. “This demonstration proves that hybrid quantum-classical feedback can now run deterministically on production hardware, setting the stage for the next generation of scalable, error-corrected systems.”
NVQLink provides the interface between quantum hardware and AI supercomputing, enabling deterministic, low-latency communication between quantum control systems and GPU servers. QM’s OPX platform implements this integration, allowing real-time execution of quantum-classical programs.
In a demonstration to be showcased at NVIDIA GTC DC, QM’s system executes a full loop between the NVIDIA CUDA-Q platform and NVIDIA CUDA: measuring qubits, transferring data to the GPU for classical computation, and returning results in real time – all within a single executable.
DGX Quantum, launched by NVIDIA and Quantum Machines in 2023, delivered the foundation for µs-latency controller-GPU integration, paving the way for NVQLink. The NVQLink architecture adds new device calls and API layers. Customers using DGX Quantum can upgrade to NVQLink without any hardware changes, preserving their current systems while gaining access to the enhanced real-time performance and new CUDA-Q capabilities of NVQLink.
“The NVQLink architecture unites quantum processors and control systems with NVIDIA AI supercomputing, delivering a powerful platform that enables builders to overcome the challenges of integrating and scaling quantum hardware,” said Tim Costa, General Manager for Quantum at NVIDIA. “Quantum Machines has already delivered pioneering work on building hybrid quantum-GPU systems, and with NVQLink they will be accelerating the next generation of quantum processors.”
By supporting the NVQLink specification, Quantum Machines joins a growing group of ecosystem partners working to standardize the integration of quantum processors and AI supercomputing resources. This open, collaborative approach paves the way toward utility-scale quantum computers capable of solving real-world problems.
