NVIDIA Ising debuts as open quantum AI platform

NVIDIA (NVDA) has introduced NVIDIA Ising, described as the world’s first family of open source quantum AI models, aimed at helping researchers and enterprises build quantum processors capable of running useful applications. The launch was announced on April 14, 2026.

Named after a landmark mathematical model used to understand complex physical systems, the Ising family focuses on two central challenges in hybrid quantum-classical systems: quantum processor calibration and quantum error correction. NVIDIA positions AI as essential for turning today’s fragile qubits into scalable, reliable quantum-GPU systems.

According to analyst firm Resonance, the quantum computing market is expected to surpass $11 billion in 2030, with progress in areas such as error correction and scalability seen as critical to that trajectory. NVIDIA states that Ising is intended to address these engineering hurdles by providing high-performance, customizable AI models, tools and data.

NVIDIA Ising includes two main components: Ising Calibration and Ising Decoding. Ising Calibration is described as a vision language model that can rapidly interpret and react to measurements from quantum processors, enabling AI agents to automate continuous calibration and reduce calibration time from days to hours.

Ising Decoding consists of two variants of a 3D convolutional neural network model, optimized separately for speed and for accuracy. NVIDIA reports that these decoding models deliver up to 2.5 times faster performance and three times higher accuracy than pyMatching, identified as the current open source industry standard for quantum error-correction decoding.

Collectively, the Ising models are presented as delivering leading AI-based quantum processor calibration capabilities and quantum error-correction decoding that is significantly faster and more accurate than traditional approaches. This is intended to let researchers tackle larger and more complex quantum problems while maintaining control over data and infrastructure through open models.

NVIDIA highlights broad adoption of Ising across quantum companies, research labs and academic institutions. Ising Calibration is already in use at organizations including Atom Computing, Academia Sinica, EeroQ, Conductor Quantum, Fermi National Accelerator Laboratory, Harvard John A. Paulson School of Engineering and Applied Sciences, Infleqtion, IonQ (IONQ), IQM Quantum Computers, Lawrence Berkeley National Laboratory’s Advanced Quantum Testbed, Q-CTRL and the U.K. National Physical Laboratory.

Ising Decoding is being deployed by Cornell University, EdenCode, Infleqtion, IQM Quantum Computers, Quantum Elements, Sandia National Laboratories, SEEQC, University of California San Diego, UC Santa Barbara, University of Chicago, University of Southern California and Yonsei University. These users are applying the models for quantum processor development and error-correction research.

NVIDIA is also providing a cookbook of quantum computing workflows, training data and NVIDIA NIM microservices. The company says these resources are designed to help developers fine-tune Ising for specific hardware architectures and use cases, including the option to run models locally to protect proprietary data.

Ising is positioned as part of NVIDIA’s broader quantum and AI portfolio. The models complement the NVIDIA CUDA-Q software platform for hybrid quantum-classical computing and integrate with the NVIDIA NVQLink QPU-GPU hardware interconnect to support real-time control and quantum error correction.

NVIDIA notes that Ising joins other open model families such as NVIDIA Nemotron for agentic systems, NVIDIA Cosmos for physical AI, NVIDIA Alpamayo for autonomous vehicles, NVIDIA Isaac GR00T for robotics and NVIDIA BioNeMo for biomedical research. These models and associated data and frameworks are available via GitHub, Hugging Face and build.nvidia.com.

On the same date, IQM Quantum Computers announced AI-driven agentic calibration that uses NVIDIA Ising as a core element. The company describes this as an automated approach to tuning quantum systems, aimed at making quantum computing operationally viable for enterprises, AI factories and high-performance computing data centres.

IQM’s architecture emphasises parallelism. Visual AI agents inspect calibration results across multiple qubits simultaneously at each stage, instead of sequentially. IQM argues that as quantum processors scale and interaction channels grow non-linearly, such parallel inspection can keep pace where traditional sequential calibration cannot.

The company positions agentic calibration as a response to the global scarcity of quantum engineering talent and the operational challenges of maintaining quantum systems. By integrating AI agents directly into its existing calibration infrastructure and fine-tuning NVIDIA Ising models for quantum tasks, IQM reports that its systems can self-optimise, sustain high algorithmic efficiency and operate consistently at high fidelity.

IQM states that this approach is intended to remove manual bottlenecks, increase uptime and deliver accurate computational results faster, without requiring on-site quantum expertise. The solution is described as open and transparent, reflecting its foundation on the NVIDIA Ising open model family.