Turing Quantum Links Photonic Chip Supply Chain and Hybrid Computing Center

At the event, founder Jin Xianmin, who also heads the Wuxi Photonic Chip Research Institute at Shanghai Jiao Tong University, signed the chip cooperation with material supplier TianTong, design-software provider Luceda and the institute's manufacturing arm. The partners aim to close gaps across materials, design tools, manufacturing and application, according to the report. The FAIC, founded in 2022 by Turing Quantum, MetaX, UCloud and NeuroXess, expanded from four founding members to dozens, adding embodied-AI company Fourier Intelligence. The Greater Bay Area center is being built with Asia Computing and Zhicheng Suanwang and is intended to integrate QPUs with CPUs and GPUs so quantum resources can be scheduled alongside conventional computing.

Jin said usable hardware alone does not complete industrialization. Quantum computing capacity must be recognized by scheduling platforms and opened to industry users, he said, adding that a single hardware breakthrough cannot close the loop without upstream and downstream support. The alliance and the center are meant to lower adaptation barriers among technology providers and turn quantum computing from a laboratory demonstration into a purchasable resource, the report said.

Among quantum routes, photonic and superconducting systems are the main ones that have publicly met the conditions of being manufacturable, scalable and having demonstrated quantum advantage, according to the report. China's Zuchongzhi No.3, a 105-qubit superconducting prototype, was 15 orders of magnitude faster than the fastest supercomputer at the time on a random circuit sampling task in 2025. Jiuzhang No.4, a programmable photonic prototype, manipulated and detected 3,050 photons in 2026 and was 10^54 times faster than the fastest supercomputer on Gaussian boson sampling. PsiQuantum's Omega photonic chipset is made on GlobalFoundries' 45nm line, with 99.98% single-qubit state preparation and measurement fidelity and 99.72% chip-to-chip interconnect fidelity, the report said.

Turing Quantum launched its third-generation TuringQ Gen3 in September 2026. The system uses an all-chip architecture based on high-speed programmable thin-film lithium niobate photonic chips and a space-time multiplexing design; its Gaussian boson sampling unit is called Zhiyuan 3. The company claims single-chip operation at the 10,000-photon level and one million modes, with chip-level quantum advantage inside a defined loss boundary. It says transmission loss on the chip is 0.03 dB/cm, key single-device insertion loss is below 0.1 dB, fiber-to-chip coupling loss is 0.9 dB per facet and wafer-level quantum light-source spectral purity exceeds 99%. From the second-generation Gen2, Turing Quantum's photonic machines have used standard IDC racks and run at room temperature, so customers do not need a special environment, Jin said.

The Wuxi institute has built a domestic pilot line for photonic chips that forms a wafer-level closed loop of design, manufacturing, testing, feedback and iteration, shortening chip iteration to two to three weeks, according to the report. Turing Quantum is pursuing hardware delivery, computing services and quantum-intelligent computing centers, and has cooperation in biomedicine and finance. Jin compared photonic quantum computing to a hammer whose commercialization depends on finding the right nails.

Industry attention is shifting from physical qubits to logical qubits. CICC said in a 2026 analysis that quantum computing remains in an early stage between noisy intermediate-scale quantum devices and fault-tolerant machines, with physical qubit count, gate fidelity, error-correction thresholds and system stability still uncertain, the report said. High-performance photonic materials and key components, wafer-level yield, loss control in complex systems and large-scale error correction remain unresolved. The main competition has moved to production yield, long-term system stability, software ecosystems and customers' willingness to pay. The report said photonic quantum computing has moved beyond principle demonstrations but still faces manufacturing, engineering and application hurdles before broad commercial use.