NVIDIA Releases First Open AI Models for Quantum Computing — Ising Cuts Calibration Time from Days to Hours
NVIDIA released Ising, the first open-source family of AI models for quantum computing, targeting the two biggest operational bottlenecks: QPU calibration and error correction. The models are already deployed at Harvard, Fermi Lab, and Lawrence Berkeley National Lab.
Original sourceNVIDIA has launched Ising, a family of open-source AI models designed specifically for quantum computing infrastructure — a first for the industry. Released on April 14, 2026 under Apache 2.0, the models address two of the most labor-intensive aspects of running quantum processors at scale: calibration and error correction.
**Ising Calibration** is a 35-billion-parameter vision-language model that reads experimental measurements from quantum processing units (QPUs) and recommends the tuning adjustments needed to keep them performing optimally. Without AI assistance, this process requires teams of specialized quantum engineers and typically takes days to complete per calibration cycle. NVIDIA claims the model reduces this to hours.
**Ising Decoding** comes in two variants (0.9M and 1.8M parameter 3D convolutional neural networks) for quantum error correction — one of the fundamental unsolved challenges in making quantum computers practically useful. The models deliver up to 2.5x faster and 3x more accurate decoding than current industry-standard tools.
The announcement sent quantum computing stocks sharply higher: IonQ gained over 20% on the day. Jensen Huang framed the release as foundational: "AI is essential to making quantum computing practical." NVIDIA has been quietly investing in quantum infrastructure tooling for years, and Ising represents the company's first major public commitment to the space.
All models are available on GitHub, Hugging Face, and build.nvidia.com. Early institutional adopters include Harvard John A. Paulson School of Engineering, Fermi National Accelerator Laboratory, UC San Diego, Lawrence Berkeley National Lab's Advanced Quantum Testbed, and the U.K. National Physical Laboratory.
Panel Takes
The Builder
Developer Perspective
“Apache 2.0 licensing on a 35B model for a niche but critical use case is a smart move. NVIDIA is positioning itself as the default infrastructure layer for quantum computing the same way it dominated traditional ML. The calibration model alone could save quantum hardware teams thousands of engineer-hours.”
The Skeptic
Reality Check
“The quantum computing market has been 'five years away' for twenty years. NVIDIA is smart to position early, but investors driving IonQ up 20% on an ancillary tooling release are getting ahead of themselves. Useful fault-tolerant quantum computers are still a decade away at minimum — this is infrastructure for infrastructure.”
The Futurist
Big Picture
“AI solving quantum's calibration problem is a meta-moment: artificial intelligence removing the human bottleneck from building the next generation of computing hardware. When quantum scales, the combination of AI + quantum will be transformative in ways we can barely model today. Ising is the first brick.”