Approaching.AI has completed a Pre-A round financing of several hundred million yuan

EO News on May 20: Recently, AI Token production service provider Approaching.AI (趋境科技) announced the completion of a Pre-A round financing worth hundreds of millions of yuan. This round of financing was jointly led by Xinglian Capital (星连资本) and Huakong Fund / T-Capital (华控基金), with Honghui Capital (弘晖资本), Top Resource Energy Co., Ltd. / Tianhao Energy (天壕能源), Shangshi Capital (尚势资本), Tianjin Ren’ai Hongsheng (天津仁爱弘盛), Hangzhou Fucheng (杭州福成), and other institutions participating as followers. Existing shareholder GL Ventures (高瓴创投) continued to increase its investment.

Upon completion of this round of financing, Approaching.AI (趋境科技) will continue to increase investment in the high-efficiency AI Token production service platform (Approaching AI Token as a Service, ATaaS), focusing on advancing computing power reserves and underlying inference system construction. It will continue to deliver model output capabilities featuring low latency, high throughput, stable structured output, reliable function calling, and predictable service quality, further enhancing the capability for large-scale supply of high-quality Tokens for enterprise production environments.

Approaching.AI (趋境科技): A core supplier of high-quality hierarchical Tokens in China, with a daily ATaaS call volume of nearly one trillion.

As large model applications enter enterprise production environments, the evaluation criteria for AI inference infrastructure are changing: enterprises no longer focus solely on computing power scale, the number of models, and interface richness, but place greater importance on whether every call can complete business delivery stably, efficiently, and predictably. At this stage, the core competitiveness of inference services is shifting towards Shifting from "Providing Models" to "Producing High-Quality Tokens". Time to first Token, Tokens per second, stability of structured output, reliability of function calling, and predictability of service quality under high concurrency scenarios are becoming important metrics for enterprises when choosing AI infrastructure.

Approaching.AI (趋境科技) believes that Token is no longer just the basic unit of input and output for large models, but a key production factor connecting model capabilities, system performance, service stability, and cost efficiency. Based on this judgment, the company proposes the industrial concept of Token as a Service (TaaS) and has built the high-efficiency AI Token production service platform ATaaS. Compared to traditional MaaS which focuses on model invocation and management, ATaaS focuses more on the delivery of inference efficiency in enterprise-level production scenarios, helping enterprises obtain scalable, operable, and high-quality Token production capabilities.

In terms of model strategy, Approaching.AI (趋境科技) insists on "Fewer models, deep optimization" The route is not to broadly support hundreds of models, but to focus on a few high-productivity models, continuously optimizing output quality, inference efficiency, TTFT stability, and TPS performance around real enterprise scenarios. For enterprise customers, the number of models does not directly equate to productivity; what truly matters is whether every call can stably support business results.

In terms of system capabilities, Approaching.AI (趋境科技) converts underlying computing power into sustainable, high-quality AI Token production capacity through capabilities such as heterogeneous computing power scheduling, cross-cluster cache sharing, inference link isolation, elastic scaling, and quality monitoring. Relying on full-link system engineering capabilities, the company can provide enterprises with more stable TTFT, 30–50 TPS high-speed output capabilities, and reliable service assurance under the premise of controllable costs.

Currently, Approaching.AI (趋境科技) has provided services to multiple enterprise-level clients such as Zhipu AI / GLM (智谱 GLM) and Moonshot AI / Kimi (月之暗面 Kimi) through the ATaaS platform, with the platform processing nearly one trillion Tokens daily. After long-term verification in high-complexity and high-concurrency business scenarios, the company has established core capabilities for large-scale inference delivery.

Commercial drive and technology foundation jointly safeguard the strategic development of ATaaS

TaaS is not an ordinary application layer product, but a systematic capability oriented towards the full AI inference chain. It requires the team to understand enterprise customer needs, industrial resources, capital pathways, and commercialization rhythms, as well as possess long-term accumulation in underlying architecture fields such as computing, storage, scheduling, caching, and inference systems. The core team of Approaching.AI (趋境科技) possesses both commercial implementation capabilities and deep technical R&D, laying the foundation for ATaaS to move from a cutting-edge concept to large-scale deployment by top-tier clients within two years of its establishment.

At the business and operational level, Approaching.AI (趋境科技) has formed organizational capabilities that synergistically advance technology productization and commercial capitalization. Founder and CEO Ai Zhiyuan holds a PhD in Computer Science from Tsinghua University, possessing both systematic research capabilities and commercialization experience from major tech companies. He spearheaded the proposal of the TaaS industry logic and has driven ATaaS to evolve from a technology platform into enterprise-level production services. President Dr. Wu Wenjie holds a PhD in Finance and a CFA qualification, with an executive background in top-tier industrial and capital institutions. She has led investments and mergers for dozens of benchmark enterprises and comprehensively oversees the company's strategy, internal control, and global operations. Chairman Ren Xuyang is a veteran of Baidu's early founding team and led the establishment of companies such as iQIYI (爱奇艺), Yidian Zixun / Yidian Information (一点资讯), Haizhi / Percent Corporation (海致), and News Break. He supports the company's development in terms of industry judgment, organizational building, capital synergy, and ecological resource integration.

In terms of technology and research, Approaching.AI (趋境科技) is backed by over twenty years of technological accumulation from the Institute of High Performance Computing at Tsinghua University, and has completed the process of capitalizing and increasing shares with relevant technological achievements from Tsinghua University. This batch of technological achievements was developed over the long term by research teams including Academician Zheng Weimin, Professor Wu Yongwei, and Associate Professor Zhang Mingxing, covering key areas such as high-performance computing, parallel and distributed systems, storage systems, intelligent computing systems, and large model inference infrastructure. The injection of these achievements marks that the industry-university-research collaboration between the company and the Tsinghua research team in the field of AI infrastructure has entered a substantive implementation phase.

Among them, Academician Zheng Weimin, Chief Scientific Advisor of Approaching.AI (趋境科技), laid the academic foundation in the field of high-performance computing at Tsinghua; Professor Wu Yongwei, Chief Scientist of Approaching.AI (趋境科技), has long been deeply engaged in distributed and storage systems and has won multiple national-level science and technology awards; Associate Professor Zhang Mingxing focuses on large model inference architectures, and the open-source projects he led, such as KTransformers and Mooncake, have been widely applied in the industry. Relying on the equity participation of core technological achievements from Tsinghua and the continuous support of a top-tier scientific research team, Approaching.AI (趋境科技) has established barriers in system engineering and research transformation oriented towards AI inference infrastructure.

This technological accumulation has also been verified in the open-source ecosystem. KTransformers, an open-source project led by Approaching.AI (趋境科技) and a Tsinghua team, is the world's first edge heterogeneous inference framework. Its GitHub Star count has surpassed 17k, and it has become the first recommended inference engine for top-tier large models such as GLM, Kimi, MiniMax, and Qwen. In the direction of distributed inference, Approaching.AI (趋境科技) has co-built the open-source project Mooncake with industry-academia-research institutions including Tsinghua University, Moonshot AI / Kimi (月之暗面 Kimi), 9#AIsoft, Alibaba Cloud (阿里云), and Ant Group (蚂蚁集团). Approaching.AI (趋境科技) technical experts and Tsinghua University Ph.D. Yang Ke, among other members, served as core contributors, deeply participating in the implementation of multiple key technologies and architecture construction. Furthermore, Approaching.AI (趋境科技) actively participates in technical contributions to global inference communities such as SGLang, vLLM, and NVIDIA Dynamo, continuously promoting the development of the open ecosystem for AI inference infrastructure.

The commercial team's grasp of industry demands, customer scenarios, and capital pathways, combined with the technical team's long-term accumulation in the fields of high-performance computing, distributed systems, and large model inference infrastructure, has enabled Approaching.AI (趋境科技) to possess complete capabilities ranging from underlying system R&D to enterprise-level large-scale delivery. With the continuous evolution of ATaaS, this composite team structure will continue to support the company in enhancing its capabilities for the large-scale production and delivery of high-quality Tokens.

Zhang Yang, Chairman of Huakong Fund / T-Capital (华控基金), stated:

With the rapid improvement in the capabilities of domestic large models and the comprehensive explosion of application demands, the massive demand for Tokens is reshaping the computing power industry chain. Huakong Fund / T-Capital (华控基金) firmly believes that the AI Infra industry, capable of providing high-quality Tokens on a large scale and stably, will become the key infrastructure for the booming development of the AI industry, possessing vast market space and extremely high investment value. The Approaching.AI (趋境科技) team originates from the Institute of High Performance Computing at Tsinghua University, possessing profound research heritage and solid technology. They have successfully broken through the underlying computing power islands and have gained high recognition from the upstream and downstream ecosystems. Huakong Fund / T-Capital (华控基金) will rely on its rich AI industry ecosystem resources to fully support Approaching.AI's (趋境科技) subsequent financing and listing process, accompanying the enterprise to grow into a globally leading new-generation intelligent computing "Token factory".

Li Wenjue, a partner at Xinglian Capital (星连资本), stated:

Approaching.AI (趋境科技) has demonstrated exceptional technical depth and engineering capabilities in the field of AI infrastructure, particularly leading the world in Token production efficiency. This round of investment values the systematic breakthroughs of its ATaaS platform, as well as the company's ability to rapidly translate top-tier academic achievements into large-scale commercial implementations. With the widespread adoption of AI Agent applications and the surge in Token demand, enterprises that can efficiently convert computing power into intelligent output will become the core of industrial competition. Xinglian Capital (星连资本) believes that whoever can synergistically optimize control, scheduling, runtime, memory management, and model structure in the future is more likely to gain the discourse power in the next generation of AI infrastructure. Approaching.AI (趋境科技) possesses both a top-tier technical background from Tsinghua University and rich commercial experience, giving investors full confidence in its sustained growth and market leadership in the AI Infra track, and firmly supporting it in building an industry benchmark for efficient AI Token production.