NVIDIA launches world's first open-source quantum AI model to help develop quantum chips

Last night.NVIDIA,Announcing the world's first open source quantum AI modelNVIDIA IsingThe company's goal is to accelerate progress toward practical quantum computers by helping researchers and businesses build quantum chips that can run practical applications.

According to NVIDIA, the NVIDIA Ising open model family offers the world's best AI-based quantum chip calibration capabilities, as well as faster-than-conventional methods of2.5 timesHigh accuracy3 timesThe quantum error correction decoding function of the

Major breakthroughs in quantum processor calibration and quantum error correction are needed to realize large-scale practical quantum applications and to build hybrid quantum classical systems.AI is the key to transforming today's quantum processors into large-scale, reliable computers.

The NVIDIA Ising series is named after a landmark mathematical model. The model dramatically simplifies the understanding of complex physical systems, providing high-performance, scalable AI tools for quantum error correction and calibration.

Jen-Hsun Huang, Founder and CEO of NVIDIA, talked about “AI is critical to enabling the practicality of quantum computing. With the Ising model, AI will become the control plane - the operating system for quantum machines - transforming fragile quantum bits into scalable and reliable quantum GPU systems.”

NVIDIA Ising contains advanced customizable models, tools and data to accelerate quantum processors:

(1) Ising calibration: a pre-trained visual language model with 35 billion parameters that quickly interprets and responds to measurement data from a quantum chip. This allows the AIintelligent bodyContinuous calibration can be automated, reducing the time required from days to hours.

(2) Ising decoding: two 3D CNN models for pre-decoding, optimized for speed or accuracy, respectively, and for real-time decoding for quantum error correction, with 900,000 and 1.8 million parameters, respectively).Ising decoding models are 2.5 times faster and 3 times more accurate than the current open-source industry standard, pyMatching.

• Paper Address:https://research.nvidia.com/publication/2026-04_fast-ai-based-pre-decoders-surface-codes
• GitHub project address:https://github.com/nvidia/ising-decoding
• Hugging Face project address:https://huggingface.co/collections/nvidia/nvidia-ising

Ising calibrations have been used by Atom Computing, Academia Sinica, EeroQ, Conductor Quantum, Fermi National Accelerator Laboratory, Harvard's John Paulson School of Engineering and Applied Sciences, Infleqtion, IonQ, IQM Quantum Computer, Lawrence Berkeley National Laboratory's Advanced Quantum Testbed, Q- CTRL and the UK National Physical Laboratory (NPL), among other organizations.

Ising decoding technology is currently deployed by Cornell University, EdenCode, Infleqtion, IQM Quantum Computer Corporation, Quantum Elements, Sandia National Laboratories, SEEQC, UC San Diego, UC Santa Barbara, University of Chicago, University of Southern California, and Yonsei University.

In addition, NVIDIA provides a set of how-to guides for quantum computing workflows and training data, as well as NVIDIA NIM microservices that enable developers to fine-tune models for specific hardware architectures and use cases with minimal setup. The models can also run locally on the researcher's system, thereby protecting proprietary data.

NVIDIA Ising works with the NVIDIA CUDA-Q software platform to enable hybrid quantum-classical computing and integrates with the NVIDIA NVQLink QPU-GPU hardware interconnect to enable real-time control and quantum error correction, providing researchers and developers with the full suite of tools they need to transform today's quantum bits into tomorrow's accelerated quantum supercomputers.

The quantum computing market size is expected to exceed $11 billion by 2030, according to analyst firm Resonance. This growth trajectory is highly dependent on continued progress in addressing key engineering challenges, such as quantum error correction and scalability.