Nvidia’s gift for Global Quantum Day on April 14, 2026 is Ising, a family of open-source AI models designed to tackle the number one problem in quantum computing: noise. By automating calibration and error correction, Nvidia hopes to turn today’s “noisy” qubits into stable and usable systems.
Quantum processors have been sensitive to even the slightest changes in temperature or magnetic fields, leading to errors. Currently, the best processors make an error roughly every 10^3 operations, but the goal is to reach a threshold of less than 10^-12 for commercial usability, which is where Ising comes in.
Ising Calibration automates the processor calibration process, saving time by interpreting experiment results and adjusting parameters in real-time. According to Nvidia, this model outperforms industry leaders on the new QCalEval benchmark.
Ising Decoding focuses on error correction during execution. Nvidia offers two optimized 3D CNN neural network variants for this task, with one model being 2.25 times faster and more precise. This high-speed decoding is crucial for scaling and managing millions of qubits, with the code available under the Apache-2.0 license on GitHub.
The strategic announcement caused a significant impact on the market, with Nvidia’s software providing reassurance to investors about stabilizing existing machines. Stocks surged for IonQ, Rigetti Computing, and D-Wave Quantum in response to Nvidia’s software release.
Institutions like Harvard, Chicago, and national labs like Fermilab have already integrated Ising into their workflows. For companies like IQM, this “agent-based calibration” makes quantum systems usable in the enterprise without the need for an onsite army of doctorate holders to monitor the machines.
