Tsinghua-linked quantum AI startup Qingxing Yishi raises nearly 100M yuan in two rounds within three months

September 24 — Tsinghua-affiliated quantum AI company Qingxing Yigou (ClearSight Heterogeneous Computing) announced the completion of its Series A+ funding round. Investors include Jingkai Capital, Anhui Hi-Tech Investment, Daode Investment, Lishi Investment, Zhongzi Fund, Xuhui Capital, Senlan Group, and the founder of Jingdong Group.

Combined with its earlier Series A round, the company has completed two financing rounds within three months, raising a total of nearly 100 million yuan.

Company Background

Founded in 2021 by Yu Teng, an alumnus of Tsinghua University's Department of Computer Science, Qingxing Yigou initially focused on parallel heterogeneous computing and automatic parallel compilation. In 2022, the company completed three consecutive funding rounds within six months.

In 2025, the team achieved mathematical validation of a high-density intelligent construction method based on quantum computing principles. Yu Teng subsequently halted the company's traditional AI business and shifted R&D focus to quantum AI, relocating operations to Shanghai's Xuhui District.

Key Product Launch

In June of this year, Qingxing Yigou released RiverONE, China's first visual language model reconstructed on simulated quantum computing power.

The model uses simulated quantum computing to generate parameters during the construction phase. After training, it performs inference on classical GPUs without requiring real-time quantum computer involvement.

According to third-party test reports, RiverONE—with only 1.9 billion parameters—achieves at least 95% of the performance of NVIDIA's 350-billion-parameter Ising Calibration 1 model in quantum calibration chart understanding tests.

The product has already been integrated with multiple chip and computing power vendors.

Industry Partnerships

• 2023: Qingxing Yigou became an AMD AI system supplier.
• Biren Technology: Completed adaptation testing of mainstream quantum computing frameworks (Cirq, PennyLane, Qiskit) on its GPUs, with no abnormal interruptions during 48-hour stability tests.
• Muxi Technology: Adapted RiverONE on the vLLM inference framework using its Xiyun C-series GPUs.
• Taichu Electronics and Loongson: Completed Day0 adaptation of Qingxing Yigou's full suite of quantum-inspired models and operators.

Commercial Application

A leading domestic optical lens manufacturer, Senlan Optics, introduced Qingxing Yigou's quantum optimization model. As a result, precision processing of optical raw materials shifted from trial-and-error to one-time scanning, improving overall efficiency by 20% to 30%.

Yu Teng estimates that each high-precision optical processing machine thus generates over one million yuan in additional value.

Strategic Collaboration

In August, Qingxing Yigou signed a strategic cooperation agreement with quantum computing company Taiyi Liangsheng to jointly develop a visual language model for debugging neutral-atom quantum computers.

Driving Force Behind Rapid Funding

The company's ability to close two funding rounds in three months is driven by the successful deployment of its quantum-generated parameter methodology on classical GPUs.

This approach bypasses the bottleneck of immature quantum hardware, allowing quantum-inspired models to be deployed directly on existing computing infrastructure.

The adaptation records with AMD, Biren, Muxi, Taichu, and Loongson serve as intermediate evidence that this route is moving from the lab to production lines.

Outlook

Current validation is concentrated in two narrow scenarios: quantum calibration and optical processing. As the technology expands into broader industrial applications, engineering delivery capabilities will become more critical than model performance metrics alone.