NVIDIA has announced the release of NVIDIA Ising, the world's first family of open-source AI models specifically designed to solve the critical challenges of quantum computing. By focusing on quantum processor calibration and error correction, NVIDIA aims to accelerate the transition from experimental qubits to scalable, high-performance quantum-GPU supercomputers.

The "Control Plane" for Quantum Machines

As CEO Jensen Huang stated, "AI is the control plane—the operating system of quantum machines." The Ising model family provides high-performance, scalable tools that address two main bottlenecks in quantum development:

1. Ising Calibration: A vision language model that automates the continuous calibration of quantum processors. It can interpret measurements and react in real-time, reducing calibration times from days to just a few hours.

2. Ising Decoding: A specialized 3D convolutional neural network (CNN) optimized for quantum error correction. It is reportedly up to 2.5x faster and 3x more accurate than pyMatching, the current industry standard.

Broad Ecosystem Adoption

A massive coalition of research labs and enterprises are already adopting Ising, including:

• National Labs: Lawrence Berkeley National Laboratory (Advanced Quantum Testbed), Fermi National Accelerator Laboratory, and Sandia National Laboratories.

• Academic Institutions: Harvard, Cornell, University of Chicago, Yonsei University, and UC Santa Barbara.

• Industry Leaders: IonQ, Atom Computing, IQM Quantum Computers, Infleqtion, and Q-CTRL.

Integrated Software & Hardware Stack

NVIDIA is providing more than just models. The Ising family integrates seamlessly with:

• CUDA-Q™: NVIDIA’s open-source software platform for hybrid quantum-classical computing.

• NVQLink™: A specialized hardware interconnect for real-time control between QPUs and GPUs.

• NVIDIA NIM™: Microservices that allow developers to fine-tune Ising models for their specific hardware architectures with minimal setup.

Local Execution for Proprietary Data

Crucially, these models can run locally on a researcher’s own systems. This ensures that proprietary quantum research data never has to leave the facility, addressing the privacy concerns of both national labs and private enterprises.

The Ising launch marks a pivotal moment where AI and high-performance computing (HPC) converge to make the "quantum advantage" a practical reality. Developers can access the models, data, and workflows today via GitHub and Hugging Face.