NVIDIA launches Ising, open AI models for quantum computing
NVIDIA has released Ising, the world's first family of open-source AI models designed to accelerate quantum processor calibration and error correction, delivering up to 2.5x faster and 3x more accurate decoding than the current open-source standard.
Score breakdown
NVIDIA announced Ising, its first open-source family of AI models targeting two of quantum computing's hardest engineering problems: processor calibration and error correction. The family includes `Ising Calibration`, a vision language model that uses AI agents to reduce calibration time from days to hours, and `Ising Decoding`, a 3D convolutional neural network offered in speed- and accuracy-optimized variants that outperforms `pyMatching` — the current open-source industry standard — by up to 2.5x in speed and 3x in accuracy. A broad ecosystem of institutions has already adopted the models, including Fermi National Accelerator Laboratory, Harvard, IonQ, and Sandia National Laboratories.
- 01NVIDIA launched Ising on April 14, 2026, calling it the world's first family of open-source AI models for quantum computing.
- 02Ising Decoding is up to 2.5x faster and 3x more accurate than pyMatching, the current open-source industry standard for quantum error-correction decoding.
- 03Ising Calibration is a vision language model that reduces quantum processor calibration time from days to hours using AI agents.
NVIDIA introduced Ising, described as the world's first family of open-source AI models for quantum computing, on April 14, 2026. The family is named after a landmark mathematical model of complex physical systems and targets two core bottlenecks in building practical quantum hardware: processor calibration and error-correction decoding. CEO Jensen Huang framed AI as "the control plane — the operating system of quantum machines," with the goal of transforming fragile qubits into scalable, reliable quantum-GPU systems.
`Ising Calibration` is a vision language model that enables AI agents to continuously interpret measurements from quantum processors and automate calibration, cutting the process from days down to hours.
The Ising family comprises two main components. `Ising Calibration` is a vision language model that enables AI agents to continuously interpret measurements from quantum processors and automate calibration, cutting the process from days down to hours. `Ising Decoding` offers two variants of a 3D convolutional neural network — one optimized for speed, one for accuracy — for real-time quantum error-correction decoding; NVIDIA claims both variants beat `pyMatching`, the prevailing open-source standard, by up to 2.5x in speed and up to 3x in accuracy. The quantum computing market is projected to surpass $11 billion by 2030, according to analyst firm Resonance.
Adoption is already underway across a wide range of organizations. `Ising Calibration` users include Atom Computing, Academia Sinica, EeroQ, IonQ, IQM Quantum Computers, Q-CTRL, Fermi National Accelerator Laboratory, Harvard's John A. Paulson School of Engineering and Applied Sciences, and the U.K. National Physical Laboratory. `Ising Decoding` is being deployed by Cornell University, Sandia National Laboratories, SEEQC, the University of Chicago, UC San Diego, UC Santa Barbara, the University of Southern California, and Yonsei University, among others.
Key facts
- 01NVIDIA launched Ising on April 14, 2026, calling it the world's first family of open-source AI models for quantum computing.
- 02Ising Decoding is up to 2.5x faster and 3x more accurate than pyMatching, the current open-source industry standard for quantum error-correction decoding.
- 03Ising Calibration is a vision language model that reduces quantum processor calibration time from days to hours using AI agents.
- 04Ising Decoding offers two variants of a 3D convolutional neural network — one optimized for speed and one for accuracy.
- 05The quantum computing market is projected to surpass $11 billion in 2030, per analyst firm Resonance.