NVIDIA opens NVQLink for hybrid quantum computing
NVIDIA released its NVQLink interconnect architecture via a new real-time API, bridging the gap between GPUs and Quantum Processing Units. The platform delivers sub-4 microsecond latency and 400 Gb/s throughput, enabling real-time quantum error correction and scaling hybrid supercomputers within the CUDA-Q ecosystem.
NVQLink standardizes the interface between classical and quantum hardware, mirroring NVLink’s impact on AI data centers. Its sub-4μs latency meets the threshold for real-time quantum error correction, while integration with CUDA-Q secures NVIDIA’s role as a gatekeeper for hybrid workflows. Early adoption by over a dozen supercomputing centers and 17 QPU builders signals a significant industry shift toward NVIDIA-standardized hybrid systems as the cudaq-realtime API opens to commercial builders.
DISCOVERED
1d ago
2026-04-14
PUBLISHED
1d ago
2026-04-14
RELEVANCE
AUTHOR
Wes Roth