Author｜Ren Tianqin

Editor｜Chen Xiaoran

On April 14 (local time), NVIDIA unveiled the world's first open-source AI model suite tailored for quantum computing, NVIDIA Ising, at the GTC 2026 conference.

Calibration and Error Mitigation

Named after a renowned model in statistical physics, this toolkit directly addresses two of the most significant challenges in quantum computing's practical deployment: quantum processor calibration and quantum error mitigation.

Ising Calibration, a visual-language model boasting 35 billion parameters, can swiftly interpret measurement data from quantum processors. This capability enables AI agents to perform continuous automated calibration, compressing tasks that previously spanned days into mere hours. Ising Decoding, built on a 3D CNN architecture, offers two optimized variants—speed-focused and precision-focused—delivering up to a 2.5-fold increase in speed and up to a threefold improvement in decoding accuracy compared to the open-source benchmark pyMatching.

Jensen Huang, NVIDIA's founder and CEO, emphasized AI's pivotal role in advancing quantum computing's practicality. Through the Ising model, AI will act as the control plane for quantum computers—essentially serving as the operating system for quantum computing—transforming fragile quantum bits into scalable, highly reliable quantum GPU systems.

The model is now available under the Apache-2.0 open-source license, with related datasets and frameworks accessible on GitHub, Hugging Face, and NVIDIA's official website. Leading academic institutions, including Cornell University, the University of Chicago, the University of California, San Diego, and Yonsei University, have already adopted it.

Spurred by this announcement, quantum computing-related stocks on the U.S. stock market surged on April 14, with SEALSQ jumping 21.03%, IonQ rising 20.16%, and D-Wave Quantum gaining 15.84%. NVIDIA's stock price also soared over 6% on the same day, marking a rare 10-day winning streak.

NVIDIA's robust financial position provides the confidence to make bold investments in quantum computing.

In fiscal year 2026, the company reported annual revenue of $215.9 billion, a 65% year-over-year increase, with GAAP net profit reaching $120.067 billion. The data center business contributed nearly 90% of the revenue, while gross margins remained high at around 75%.

Strategic Vision

Despite this seemingly "flawless" financial performance, NVIDIA exhibits a strong sense of urgency. As Moore's Law nears its physical limits, the performance improvement curve for GPUs is flattening. Meanwhile, once quantum computing becomes viable, it could disrupt classical computing paradigms in specific domains like drug discovery and materials science.

After Google announced in October 2025 that it had achieved "verifiable quantum advantage," the CEO of D-Wave candidly remarked, "If I were NVIDIA, I'd be terrified." Rather than resting on its GPU laurels, NVIDIA is strategically allocating resources to the still-nascent field of quantum computing, essentially preparing for the "post-GPU era." This "AI money-printing machine," with annual revenues exceeding $200 billion, recognizes that its true competitive advantage lies not in a single hardware product but in shaping the ecological narrative of the next-generation computing paradigm.

Quantum computing has long been plagued by the "five-year curse," with repeated predictions of large-scale adoption in the next five years, only to see another five years pass without realization. The fundamental issue is that quantum bits are highly unstable—current state-of-the-art quantum processors may err once every thousand operations, while large-scale applications require error rates below one in a trillion. Calibration and error mitigation are precisely the "AI-shaped workloads" that AI is best equipped to handle.

Ising Calibration has collaborated with institutions like Fermilab and Harvard to develop QcalEval, the world's first evaluation benchmark, outperforming top proprietary models like Gemini 3.1 Pro and GPT 5.4 across six dimensions.

Ising Decoding functions as a "pre-decoder," enabling leading quantum hardware companies like Atom Computing and IonQ to integrate it directly into their platforms, facilitating the transition from 12 logical quantum bits to 30 by the end of the year.

NVIDIA's choice of a permissive open-source license and provision of NIM microservices is not merely a commercial gesture but a shrewd ecological strategy. During this chaotic period before unified standards emerge in quantum computing, NVIDIA aims to bind developers and research institutions to its toolchain by open-sourcing models, making Ising the de facto standard for future quantum systems.

A High-Stakes Endeavor

As Jensen Huang has stated, AI will become the operating system for quantum machines, and NVIDIA is attempting to run this operating system on its own GPUs.

The strategic value of Ising extends far beyond the tools themselves. In future heterogeneous computing architectures, QPUs (quantum processing units) are poised to become the third major co-processor in data centers, following CPUs and GPUs. NVIDIA's goal is to control the layer that connects them all.

The global quantum computing market is projected to surpass $11 billion by 2030. Regardless of which quantum hardware vendor ultimately prevails, all will require AI-assisted calibration and error mitigation—a role NVIDIA is positioning itself to fill as the "shovel seller" in this gold rush.

It's worth noting that commercialization may still take 5-10 years, with little substantive revenue contribution in the short term. Moreover, with multiple technical routes—such as superconducting, trapped-ion, and photonic quantum computing—still vying for dominance, it remains uncertain whether NVIDIA's selective support will align with the mainstream.

A deeper risk lies in the fact that NVIDIA is essentially preparing for a future where GPUs may be partially supplanted. If quantum computing achieves disruptive advantages in certain tasks, will today's strategic moves ensure NVIDIA remains the "indispensable" computing platform? In 2025, quantum pure-play stocks D-Wave and Rigetti surged 211% and 45%, respectively, outpacing NVIDIA's 38.8% annual gain.

Capital markets are already hedging their bets for the quantum era.

The open-sourcing of NVIDIA Ising marks a new phase of deep integration between AI and quantum computing, two cutting-edge technologies. From GPU dominator to builder of a quantum "operating system," NVIDIA is waging a war that spans two eras.

The outcome of this high-stakes gamble remains uncertain, but one thing is clear: in the chess game of next-generation computing paradigms, NVIDIA has just made a highly significant move.