NVIDIA has launched a new family of open AI models aimed at solving two of the biggest challenges in quantum computing: calibration and error correction.
Called Ising, the models are designed to help researchers and enterprises build more stable and scalable quantum systems by improving how quantum processors are tuned and how errors are detected and corrected in real time.
Quantum computers are highly sensitive systems where even small disturbances can lead to errors. Fixing these issues has been one of the biggest barriers to building machines that can handle real-world applications at scale.
NVIDIA says its Ising models deliver up to 2.5 times faster performance and three times higher accuracy in quantum error correction compared to traditional approaches, marking a step toward more reliable quantum hardware.
Fixing fragile quantum systems
The Ising family includes tools for both calibration and decoding. Calibration ensures that quantum processors are correctly tuned, while decoding helps identify and correct errors that occur during computation.
The Ising Calibration model uses a vision-language approach to interpret measurement data from quantum processors.
This allows AI agents to automate calibration processes that traditionally take days, reducing them to hours. Faster calibration cycles also allow researchers to run more experiments and improve system performance over time.
On the error correction side, Ising Decoding uses 3D convolutional neural networks to process complex quantum data in real time.
The models are optimized for both speed and accuracy, enabling faster correction of errors as they occur. This is critical for maintaining coherence in quantum systems during longer computations.
“AI is essential to making quantum computing practical,” said Jensen Huang, founder and CEO of NVIDIA. “With Ising, AI becomes the control plane — the operating system of quantum machines.”
AI meets quantum hardware
The models are being adopted by a wide range of institutions, including research labs, universities, and quantum computing companies.
Early users include Harvard University, Fermilab, Lawrence Berkeley National Laboratory, and several commercial quantum firms.
This level of adoption highlights the growing role of AI in managing quantum systems, especially as researchers work toward building hybrid quantum-classical architectures.
These systems rely on tight coordination between quantum processors and classical computing resources.
NVIDIA is also positioning Ising as part of a broader ecosystem. The models integrate with its CUDA-Q platform and NVQLink hardware interconnect, allowing real-time interaction between quantum processors and classical GPUs.
The company has made the models open, giving developers access to tools, data, and workflows to customize them for different quantum hardware setups. \This could help accelerate experimentation while allowing organizations to maintain control over their data. It also lowers barriers for smaller research teams to participate in quantum development.
The launch reflects a wider push to turn quantum computing from a research effort into a practical technology.
By improving calibration and error correction, AI models like Ising could help reduce the gap between current experimental systems and scalable quantum machines.