NVIDIA Launches Ising: Open AI Models to Supercharge Quantum Computing
NVIDIA unveils Ising, the world's first family of open-source AI models built to accelerate fault-tolerant quantum computing through AI-powered calibration and error correction.
NVIDIA has announced Ising, the world's first family of open-source AI models purpose-built to accelerate the path to useful, fault-tolerant quantum computers. Unveiled on April 14, 2026, the Ising model family targets two of the most stubborn challenges in quantum hardware: processor calibration and real-time error correction -- bottlenecks that have historically demanded enormous manual effort and slowed the pace of progress in the field.
The Ising family ships in two distinct domains. Ising Calibration is a 35-billion-parameter vision-language model that interprets measurement results from quantum processors in real time, compressing calibration work that once took days into a matter of hours. Ising Decoding, built on a 3D convolutional neural network, performs real-time quantum error correction and ships in two variants tuned for speed and accuracy respectively. Benchmarked against pyMatching -- the current open-source industry standard -- Ising Decoding delivers up to 2.5x faster throughput and 3x higher decoding accuracy, a leap researchers say is critical for scaling quantum systems to thousands of qubits.
NVIDIA CEO Jensen Huang framed the release as a step toward bridging classical and quantum computing. The models are released under an open license on Hugging Face and the NVIDIA NGC catalog, allowing academic labs and commercial quantum hardware makers to fine-tune and deploy them on their own systems. Early adopters include Fermi National Accelerator Laboratory, Harvard School of Engineering, IQM Quantum Computers, Infleqtion, and the UK National Physical Laboratory.
The launch signals a broader strategic push by NVIDIA into the quantum sector, complementing its existing Quantum Cloud initiative and the CUDA-Q programming model. Analysts note that open-sourcing Ising gives NVIDIA a way to establish GPU infrastructure as the de facto compute backbone for quantum-classical hybrid workloads -- a market expected to grow substantially as quantum hardware matures. For the quantum research community, the arrival of well-resourced, high-performance open models represents a significant acceleration in the timeline toward practical quantum advantage.
