CUDA-Q is NVIDIA's open-source platform for hybrid quantum-classical computing that provides a unified programming model for writing algorithms that seamlessly integrate GPU-accelerated classical computation with quantum processor execution. Why it matters. One of the practical challenges of hybrid quantum-classical computing is the programming model: how do you write code that naturally combines classical preprocessing on CPUs and GPUs with quantum subroutines on a QPU? CUDA-Q addresses this by providing a single framework where developers can express both classical and quantum logic, with the runtime managing the distribution across available hardware. It supports multiple quantum backends and is designed to scale from simulation on laptops to execution on real quantum processors connected to GPU clusters.
CUDA-Q is the programming layer of NVIDIA's hybrid quantum-classical stack, while NVQLink provides the underlying system architecture that connects quantum control hardware to GPU resources. How it connects. The Qblox software stack's open architecture and Python-based Qblox Scheduler are designed to integrate with third-party frameworks and programming environments. As CUDA-Q and similar hybrid platforms mature, Qblox's well-defined software interfaces (from high-level Scheduler to low-level Q1ASM) provide the hooks needed for these frameworks to target Qblox-controlled quantum hardware. Learn more about Qblox Scheduler.
