Qingxing Heterogeneous Raises A+ Round for Quantum-Inspired AI Models

Qingxing Heterogeneous Raises A+ Round for Quantum-Inspired AI Models

Qingxing Heterogeneous completed an A+ financing round, bringing its combined A and A+ funding over the past three months to nearly RMB 100 million, or about $14 million. The Tsinghua-linked company, founded in 2021 by Yu Teng, began in heterogeneous-computing software before shifting its research focus toward quantum-inspired AI. Its investors include Jingkai Capital, Anhui High-Tech Investment, Daode Investment, Lishi Investment, Zhongzi Fund and Xuhui Capital.

The company’s current commercial bet is RiverONE, a quantum-inspired vision-language model. Qingxing says it uses simulated quantum computing to generate model parameters during construction, while inference runs on conventional GPUs without a quantum computer in the loop. In a cited specialized chart-understanding test, the company said its 1.9 billion-parameter model reached at least 95% of Nvidia Ising Calibration 1 performance with less than one-tenth of the comparison model’s parameters. That claim is narrow and should not be treated as a general model-performance result, but it points to a practical route for quantum-derived methods: lowering model size and token use on hardware enterprises can deploy today.

For builders, the relevant question is not whether quantum hardware is ready, but whether the company can convert its parameter-generation approach into repeatable gains in quality, latency, cost and deployment reliability. Qingxing has already reported integration work with AMD, Biren Technology, MetaX, Loongson-based systems and game-chip company Qianhe Yibang. Investors should watch for independently validated benchmarks, durable software revenue from those integrations, and evidence that its GPU compatibility work becomes a scalable product rather than bespoke engineering.

#QuantumAI #FoundationModels #Inference #ChinaAI #Funding