Nvidia has released an open-source quantum AI model called “Ising.” Jensen Huang: AI will become the quantum operating system

Quantum computing moves toward commercialization with a major breakthrough! NVIDIA officially announced today (14th) the world’s first open-source quantum AI model family, “NVIDIA Ising,” directly tackling the toughest challenges in quantum computing: calibration and error correction. Its AI decoding performance is not only 2.5 times faster than traditional standards, but also improves accuracy by 3 times. CEO Jensen Huang boldly declared, “AI will become the operating system of quantum machines.”
(Background: StarkWare researcher Tibi’s “Quantum-Resistant Bitcoin Solution QSB”: it can withstand quantum threats without soft forks, but the per-transaction cost exceeds $75)
(Background update: The Bitcoin quantum threat is not a technical issue; Grayscale research says community consensus is the real bottleneck)

Quantum computing has long been viewed by the tech industry as the ultimate weapon in the next-generation computing revolution. However, the underlying nature of its fundamental “qubits” is extremely fragile and highly susceptible to interference from environmental noise, resulting in extremely high error rates—an obstacle that has always been the biggest stumbling block to its shift toward practical use and large-scale commercial applications.

To solve this century-old challenge, AI chip giant NVIDIA chose April 14th’s “World Quantum Day” to drop a major bombshell, officially unveiling the world’s first open-source AI model family designed specifically for quantum computing — NVIDIA Ising.

Solving calibration and error correction with two major challenges, performance leaves traditional standards far behind

NVIDIA Ising (named after the famous Lenz-Ising ferromagnetic model in statistical mechanics) is designed to provide high-performance, scalable AI tools for quantum hardware developers, researchers, and enterprises, directly targeting the two most critical challenges in building hybrid quantum-classical systems: quantum processor calibration and quantum error correction.

This open-source model family includes two core tools:

• Ising Calibration (Calibration Model): This is a powerful visual language model (VLM) that can quickly interpret the measurement results of quantum processors and respond accordingly. It enables AI agents to “automate” continuous hardware calibration tasks, cutting down the originally tedious process that took days to just within a few hours.
• Ising Decoding (Decoding Model): It includes two 3D convolutional neural network (CNN) variants optimized for either “speed” or “accuracy,” used to perform real-time quantum error correction decoding. According to NVIDIA’s data, this model is up to 2.5 times faster than the current industry open-source standard (pyMatching), and accuracy has improved by 3 times.

Jensen Huang: AI will become the “control plane” of quantum machines

For this major announcement expanding into the quantum domain, NVIDIA’s founder and CEO Jensen Huang provided high praise and a clear technical positioning:

“AI is crucial for making quantum computing practical. With Ising, AI will become the control plane — the operating system of quantum machines — turning fragile qubits into scalable and reliable quantum-GPU systems.”

NVIDIA emphasized that, through fully open-source models, developers can deploy these models to run and fine-tune on local research systems. This not only enables optimization for specific quantum hardware architectures, but also ensures absolute security of enterprise proprietary QPU data. In addition, Ising can seamlessly integrate with NVIDIA’s existing CUDA-Q software platform and NVQLink hardware interconnection technology.

Currently, NVIDIA Ising has rapidly gained widespread adoption among top global enterprises, academic institutions, and national laboratories. Including Taiwan’s Academia Sinica, Harvard University, Cornell University, the U.S. Fermi National Accelerator Laboratory, as well as quantum pioneers such as IQM Quantum Computers, Infleqtion, and Atom Computing—all have introduced Ising into their R&D lineups, jointly accelerating the arrival of next-generation quantum supercomputers.