Recently, high-quality AI Token production service provider Approaching.AI completed a Series B funding round, raising over RMB 1.3 billion; the company has cumulatively raised RMB 2.4 billion so far this year. The company states that multiple core operating metrics have achieved over tenfold growth within half a year.
This funding round was jointly led by Zhecai Innovation, Xiaoshan Capital, ICBC Investment, Zhongguancun Science City, Xinglian Capital, and Qianjiang Century City. Cloud Ventures Capital, Hunan Gaoxin Zongheng, JianTou HuaKe, Shenwan Innovation Investment, CITIC Juxin, Zhuhai Technology Industry Group, SenseTime, Wangsu Science & Technology, Gansu Shuzhi, Changjinkong, Yuanzhi Future, Qingzheng Xingzhu, Xiaoka Capital, and others participated, while existing shareholder Shangshi Capital continued to follow on.
This marks another large funding round for Approaching.AI shortly after completing its Series A round in July this year. As of October this year, the company has completed multiple consecutive funding rounds in 2026, with total funding for the year reaching RMB 2.4 billion.
The proceeds from this round will be focused on expanding high-quality AI Token production capacity, upgrading and iterating the ATaaS (Approaching.AI Token as a Service) platform, research and development of inference underlying technology and new products, as well as improving commercialization cooperation and delivery systems.
On the business side, Approaching.AI positions itself as a high-quality AI Token service provider. The company states that in the early stage of the Token market's development, the industry's supply side emphasized Token quantity over service quality, rapidly expanding supply scale through extensive volume growth; as the industry focus shifts to Agent-driven business closed-loop implementation, large models gradually carry production-grade core processes, and enterprises impose production-grade requirements on AI Token intelligence, response latency, output speed, context completeness, and service stability, leading to quality stratification in the Token market.
According to the company's definition, high-quality AI Token refers to AI Token services anchored in production-grade scenarios, providing high intelligence, low latency, high speed, strong stability, along with scalable delivery capability and controllable overall cost.
On the technology route, Approaching.AI adheres to fewer models, deep optimization, and high quality, focusing on mainstream large models with clear market demand and broad applications, and advancing engineering implementation of technologies such as full-system heterogeneous collaboration, computation-for-storage exchange, and virtual-real isomorphism.
In domestic computing power, the company has successively proposed technical solutions including domestic Prefill-Decode (PD) heterogeneous collaboration, high-performance heterogeneous KV Cache conversion, and heterogeneous computing power computation fusion, which have already been put into production in high-standard production scenarios. The company states that in real production-grade environments, it has achieved domestic chips carrying trillion-level leading large models and achieved large-scale production of high-quality AI Tokens, with overall cost-performance surpassing the world's most advanced computing power.
In the industrial ecosystem, the company has established cooperation with enterprises such as SenseTime, Parallel Technology, Hongxin Electronics, Wangsu Science & Technology, Moore Threads, MetaX, and Tianshu Zhixin.
Approaching.AI CEO Ai Zhiyuan stated that the true barrier in the computing power race is unit energy efficiency and commercial delivery capability, not scale. The company's over tenfold growth in half a year confirms the potential of the fewer models, deep optimization route. In the future, the company will use this round of funding to expand production capacity, deepen domestic PD heterogeneity, and support customers' large-scale implementation with low-cost, high-quality AI Token services.
Comments