On the afternoon of November 2, 2026, the 2026 World Internet Conference Leading Science Awards were presented in Wuzhen to 20 projects worldwide. The Jiuzhang 4.0 photonic quantum computing prototype—submitted jointly by the University of Science and Technology of China (USTC), Hefei National Laboratory, the Jinan Institute of Quantum Technology, and Jiuzhang Quantum—won an award in the basic research category, the most competitive group this year: only 2 awards were given from more than 430 submissions spanning 43 countries and regions.
The machine's specifications were previously published: 1,024 quantum squeezed-state inputs and 8,176 modes, with the first manipulation and detection of quantum states of 3,050 photons—more than a 10× increase over the 255 photons of the previous-generation Jiuzhang 3.0. For Gaussian boson sampling, generating one sample takes just 25 microseconds; by estimates based on the best classical algorithms, the world's fastest supercomputer, El Capitan, would need more than 10^42 years. The quantum advantage ratio reaches the order of 10^54, currently the strongest quantum-advantage claim internationally.
The award is an endorsement, but what really matters is the steps this machine took over four years.
Six Years, Four Generations
The evolution of the Jiuzhang series is essentially a tug-of-war against photon loss.
- 2020 – Jiuzhang: 76 photons; first quantum computational advantage in an optical system
- 2021 – Jiuzhang 2.0: 113 photons; phase-programmable
- 2023 – Jiuzhang 3.0: 255 photons; sustained lead
- 2026 – Jiuzhang 4.0: 3,050 photons; spatio-temporal hybrid encoding
- Guangming Net: "Jiuzhang 4.0 Photonic Quantum Computing Prototype Wins 2026 World Internet Conference Leading Science Award," 2026-11-02
- USTC News: "Jiuzhang 4.0 Wins World-Class Leading Science Award" (Anhui Business Daily), 2026-11-03
- Zhejiang Online: on-site report from the 2026 WIC Leading Science Awards, 2026-11-03
- *Nature* paper: Jiuzhang 4.0 photonic quantum computing prototype, 2026-05-13
In the same period, USTC's 56-qubit superconducting prototype Zuchongzhi 2.0 also succeeded, making China the only country to demonstrate quantum advantage on both photonic and superconducting platforms.
Photonics has a fatal weakness: photon loss. The larger the optical network, the more likely photons are lost en route—scale up, and effective computing power may be eaten by loss. Jiuzhang 4.0's answer is a spatio-temporal hybrid encoding architecture, in which photons interfere across both spatial and temporal dimensions: the same set of physical devices reappears in different temporal modes, so mode counts rise while physical device count is contained. As the team's CTO summarized at the ceremony: physical costs fell dramatically while system scale leapt forward—an advance from "achieving" quantum advantage to "making it scalable."
The paper was published in *Nature* on May 13 this year. The architecture keeps being cited because it proved, for the first time, that the photonic route's advantage is not just a single-purpose machine that grows ever larger—scale expansion itself has an engineering solution.
Three Boundaries to Keep in Mind
1. Advantage does not mean universality. Gaussian boson sampling is a well-defined specialized mathematical task. Jiuzhang 4.0 wins at "this one task, where a quantum machine overwhelms the strongest supercomputer." It cannot run Shor's algorithm or do general combinatorial optimization. Reading "quantum advantage" as "quantum computers comprehensively beating electronic computers" is the most common misreading of the term.
2. The 10^54 ratio is an estimate. The figure is "estimated by optimal classical algorithms" and depends on the ceiling of classical methods. Every advance in classical algorithms rewrites this ratio—the Jiuzhang team itself notes that the dynamic contest between quantum and classical is what drives the field forward.
3. The next hard battle is error correction. The team's stated directions are large-scale quantum state preparation and quantum error correction, moving toward universal fault-tolerant quantum computing. From "a sampling-specific machine" to "a fault-tolerant quantum computer" lies bosonic error-correcting codes, cluster-state preparation, and the hurdle of single-photon source brightness—each harder than the number on the trophy.
Roadmap: quantum advantage (achieved) → scalable architecture (Jiuzhang 4.0's step) → large-scale quantum state preparation → quantum error correction → universal fault-tolerant quantum computing.
Why This Award Is Worth Remembering
The Leading Science Award criteria are frontier significance, breakthrough value, and industrial impact. Jiuzhang 4.0's win signals mainstream recognition that the photonic route—which does not depend on ultra-low temperatures—has moved from "a single-point laboratory record" into "a sustainably iterating engineering system." Four generations in six years, photon count up 40×: that pace is fast by any hardware standard.
For China's quantum industry, the award has another practical effect: the supply chain Jiuzhang 4.0 pulls along—from single-photon detection to precision control and optical components—is beginning to be independently priced by capital markets. At the same day's "Light of the Internet" expo in Wuzhen, Turing Quantum's photonic quantum machine also won a "New Light" product award. Two tracks—research and industry—are advancing in parallel for photonic quantum computing.
Trophies for quantum advantage will keep changing hands; every time classical algorithms catch up, the ratio shrinks. Jiuzhang 4.0's real legacy is writing the word "scalable" into photonic quantum computing's engineering playbook—and that will not depreciate with classical progress.
---
Sources: