Qujing Technology (趋境科技) raises hundreds of millions in Pre-Series A funding to scale AI Token production infrastructure
The AMW Read
Funding update for an early-stage AI infrastructure player; confirms known TaaS trend without disrupting existing market structure.
Qujing Technology (趋境科技) raises hundreds of millions in Pre-Series A funding to scale AI Token production infrastructure
Qujing Technology, a Chinese AI infrastructure startup operating under the brand Approaching.AI, has raised several hundred million RMB (hundreds of millions of dollars) in a Pre-A funding round. The round was co-led by Xinglian Capital and Huakong Technology, with participation from Honghui Capital, Tianhao Energy, Shangshi Capital, Tianjin Renai Hongsheng, Hangzhou Fucheng, and existing investor Hillhouse Ventures. The company plans to use the capital to expand compute reserves and build out its underlying inference system, advancing its ATaaS (Approaching AI Token as a Service) platform, which currently processes nearly one trillion tokens daily for enterprise clients including Zhipu AI (智谱) and Moonshot AI's Kimi.
Why it matters: This funding exemplifies the emerging "Token as a Service" (TaaS) business model, which reframes the AI infrastructure value proposition from model access to reliable, high-quality token production at scale. Qujing's strategy — focusing on a small number of deeply optimized models rather than a broad model catalog — mirrors the inference-efficiency playbook seen in the broader AI infrastructure segment. The company's ties to Tsinghua University's high-performance computing research, including contributions to open-source inference engines KTransformers and Mooncake (co-built with Moonshot AI and Alibaba Cloud), position it as an infrastructure layer that abstracts away the complexity of distributed inference for Chinese foundation model labs. The round comes as enterprise demand for predictable latency, structured output, and function-calling reliability becomes the dominant procurement criteria, shifting the battleground from model breadth to inference quality.
An industry analyst notes: "Qujing is betting that China's enterprise AI shift will reward specialized token factories over general-purpose model hubs. Its funding success, backed by top-tier VCs and a Tsinghua research lineage, signals that the TaaS thesis is gaining traction. The challenge will be whether it can maintain cost advantages as hyperscalers like Alibaba Cloud and Tencent Cloud build competitive inference services."
#AIinfrastructure #TokenService #InferenceEfficiency #ATaaS #ChinaAI #QujingTechnology
