Qblox Achieves Sub-4 Microsecond Latency Integrating NVIDIA DGX Spark and RTX PRO 6000 Blackwell via NVQLink

Quantum control electronics provider Qblox has announced the integration of its scalable quantum system controller with NVIDIA DGX Spark and desktop workstations using NVIDIA NVQLink, achieving measured real-time roundtrip latencies below 4.0 microseconds on NVIDIA RTX PRO 6000 Blackwell GPUs and below 7.2 microseconds on NVIDIA DGX Spark.

Building on initial NVQLink architecture announcements in late 2025 and real-time API integrations with NVIDIA CUDA-Q in March 2026, the milestone validates hardware-executed, closed-loop communications between Qblox's Cluster control plane and GPU processing nodes. By embedding GPU acceleration directly into the active quantum control loop via Qblox's Q1 sequence processors and LINQ low-latency backplane, the system enables continuous, real-time qubit parameter calibration during running experiments and provides sub-4.0 microsecond transport for quantum error correction (QEC) syndrome data to GPU-based decoders.

The desktop-scale hardware setup provides researchers with a unified development environment capable of running quantum, AI, and classical simulation workloads locally via CUDA-Q, with identical control code scaling directly to GPU supercomputer clusters as physical qubit counts expand.

Review the announcement on the Qblox Newsroom here.

October 22, 2026