China’s “Quantum–Supercomputing–AI” Signal: A Smart-Computing Blueprint Beyond Quantum Chips

China’s quantum agenda is being framed as something much broader than a race to build a more powerful quantum processor.

At the Fifth CCF Quantum Computing Conference and Fifth Greater Bay Area Quantum Science Forum, held in Shenzhen from 3 to 5 August 2026, more than 2,400 people attended, while more than 200 academicians and experts presented research and industry perspectives. 23 The event’s theme—“quantum and intelligence integrated, quantum–supercomputing–AI computing in harmony”—pointed toward a combined smart-computing architecture.

That message should not be mistaken for a formally issued Chinese national plan. Instead, it was a strategic direction and set of infrastructure proposals discussed by the participating scientists, engineers and industry representatives. But the direction was clear: quantum computing, supercomputing and artificial intelligence should be developed as complementary layers of one national computing system.

What does “quantum–supercomputing–AI” mean?

The phrase describes cooperation among three types of computing power rather than a project focused solely on quantum hardware:

• AI can make quantum systems more usable by improving hardware calibration, control, design and algorithm performance.
• Supercomputers can support large-scale simulation and AI for Science, including research in materials, physics and biology.
• Future fault-tolerant quantum computers could add a new computing layer for selected scientific problems that remain difficult for classical machines.

In practical terms, this is closer to a national computing-infrastructure blueprint than to a single quantum-chip initiative. The Shenzhen event covered basic research, enabling technologies, industrial applications, standards, exhibitions and government–industry–investment matchmaking, with nearly 30 sessions held over three days. 23

AI’s role in closing quantum computing’s error gap

Quantum computing’s central engineering problem is not simply processing more qubits. It is keeping those qubits reliable enough for useful calculations.

Chinese Academy of Sciences academician Xiang Tao compared the roughly 1% error rate of current physical qubits with the far lower error levels required for demanding applications such as cryptographic codebreaking. He identified AI as an important tool for narrowing that gap. 1

At the conference, researchers described several areas in which AI could assist:

1. Chip calibration and design: AI can analyse experimental data, identify error patterns and help optimise device parameters.
2. Quantum error correction and decoding: Machine-learning methods may improve the speed and accuracy with which errors are detected and corrected.
3. Quantum-circuit compilation: AI can help map abstract algorithms more efficiently onto the constraints of real quantum hardware.
4. Measurement and control: AI can improve the real-time operation of fragile quantum systems.
5. Variational quantum algorithms: AI can accelerate parameter searches and help address “barren gradients”—situations in which gradients become too small to guide training effectively. 1

This gives AI a role well below the application layer. It could become part of the basic toolkit used to design quantum chips, operate experimental systems, compile circuits and refine algorithms. In that sense, AI is not merely a customer for quantum computing; it may also be one of the technologies needed to make quantum computing work.

From a quantum project to an integrated computing system

CAS academician Xue Qikun said that closer integration among AI, supercomputing and quantum computing is already reshaping quantum-materials research and the development of quantum algorithms. 1

That view expands the discussion from an individual machine to a wider national capability. Quantum processors require classical systems for control and data processing. AI depends on large datasets and high-performance computing. Scientific discovery often requires several kinds of computing resources to be coordinated in one workflow.

Chinese Academy of Engineering academician Gao Wen proposed a Pengcheng Laboratory-led infrastructure combining 10-gigascale computing, 10-petascale networking and 10-gigawatt-scale clean energy to provide integrated resources for national laboratories. 1

In the conference’s context, these figures represented more than a collection of hardware targets. They reflected an approach in which computing, networking and energy supply are planned together.

That is the strategic heart of the “quantum–supercomputing–AI” concept. If computing capacity is treated as a foundation for scientific, industrial and security capabilities, the competitive unit is no longer a single quantum computer. It is the broader system capable of continuously supplying computing power, data, energy, skilled personnel and real-world applications.

Supercomputers are the near-term foundation for AI for Science

CAE academician Zheng Weimin described today’s leading supercomputers as, in effect, enormous AI machines for scientific research. In this architecture, quantum computing would become a further layer of capability. 1

That creates a relatively pragmatic timeline. Quantum computing may have significant long-term potential, but supercomputing and AI for Science can already support work such as materials discovery and protein-structure prediction. 1

Research organisations can therefore use mature classical high-performance-computing and machine-learning infrastructure now, while gradually assigning suitable problems to quantum systems as those systems improve. This layered model reduces the need to wait for a general-purpose, fault-tolerant quantum computer before pursuing useful applications.

It also reflects the reality that quantum computing still faces major challenges involving error rates, error correction and scaling. AI and supercomputing can contribute immediately to algorithm development, experimental calibration and scientific-data processing while the quantum hardware continues to mature.

Materials discovery offers a concrete example

One workflow presented at the conference combined the quantum many-body principle of sigma-bond metallisation with AI-based screening. It searched a database containing more than two million materials and identified 31 candidate high-temperature superconductors, including 18 that had not previously been predicted. 1

The significance of the example is not that AI replaced physics. Rather, the workflow divided the work among complementary tools:

• physics supplied the theoretical principle and search criteria;
• AI explored a very large candidate space;
• supercomputing provided the processing and verification capacity needed to handle the data.

CAS academician Gong Xinguo also argued that AI could allow researchers to define problems in natural language and call on specialised AI “skills” for materials calculations, design and synthesis. 1 If that approach matures, the entry point for materials research could shift from writing complex computational workflows to defining questions, combining tools and validating the results.

A system-level response to the US quantum push

Gao Wen compared China’s integrated approach with the US Department of Energy’s Quantum Genesis initiative, announced in 2026. DOE materials describe the initiative as an effort to create and deploy scientifically relevant, fault-tolerant quantum computers while coordinating federal agencies, national laboratories and industry.

The comparison made in Shenzhen was strategic rather than a claim that the two countries have identical programmes. Their objectives, institutional structures and investment scopes differ. What can be compared is the direction of travel: both approaches treat quantum computing as part of a larger national science, infrastructure and industrial effort rather than as an isolated hardware project.

Is this an official Chinese national blueprint?

Strictly speaking, no. The available material presents “quantum–supercomputing–AI” as the conference theme and as a strategic assessment and set of proposals from participating academicians and experts—not as a formally promulgated national plan. 123

Even so, the discussions outlined a coherent policy pathway:

• use AI to improve quantum hardware, software and algorithms;
• use supercomputers as the near-term platform for AI for Science;
• develop quantum computing as a future layer for specific complex problems;
• plan computing, networking, clean energy and national-laboratory resources as an integrated infrastructure;
• connect universities, research institutes, companies, government agencies, users and investors to move laboratory results toward scaled applications. 237

The most important signal from Shenzhen, then, was not that quantum computing has suddenly entered a fully practical era. It was that China’s competitive narrative is broadening—from “who has the strongest quantum device?” to “who can organise quantum computing, AI, supercomputing and scientific infrastructure into the most effective system?”