NVIDIA Launches Ising Open AI Models to Accelerate Path to Useful Quantum Computers

The Control Plane for Quantum Machines

NVIDIA unveiled the world’s first open source quantum AI model family today, aiming to solve the critical challenges of processor calibration and error correction. Named NVIDIA Ising, the suite provides researchers with high-performance tools to transform fragile qubits into reliable quantum-GPU supercomputers. This release positions artificial intelligence as the essential control plane for the next generation of accelerated computing.

Breakthroughs in Calibration and Decoding

The Ising family includes specialized models tailored for the physical demands of quantum hardware. Ising Calibration uses a 35-billion-parameter Vision-Language Model to interpret raw telemetry data, reducing hardware tuning time from days to just a few hours. Meanwhile, Ising Decoding utilizes 3D Convolutional Neural Networks to perform real-time error correction, outperforming the current industry standard by 2.5x in speed and 3x in accuracy.

"AI is essential to making quantum computing practical," stated Jensen Huang, founder and CEO of NVIDIA. "With Ising, AI becomes the control plane, the operating system of quantum machines, transforming fragile qubits to scalable and reliable quantum-GPU systems."

Global Ecosystem Adoption

Global research institutions and enterprises have already begun adopting the technology to streamline their workflows. Partners such as Harvard, Fermilab, and IQM Quantum Computers are utilizing the calibration models to automate complex hardware loops. Because the models are open source, developers can run them locally to maintain total control over proprietary data and sensitive infrastructure while integrating with NVIDIA CUDA-Q for hybrid operations.

Accelerating the Path to 2030

By providing weights and training frameworks on Hugging Face and GitHub, NVIDIA is lowering the barrier to entry for quantum development. This strategic move supports a market expected to exceed $11 billion by 2030 by removing the classical computing bottlenecks that hinder scalability. These tools are now available for fine-tuning via NVIDIA NIM microservices to support diverse hardware architectures.