Who is producing the "water, electricity, and coal" of the AI era?
A round of financing has come into our view—on July 13, Qujing Technology's Series A funding came to light, significantly led by Henan Investment Group's Huirong Fund, with old shareholders including Zhenzhi Capital, Shangshi Capital, Xinglian Capital, Shanghai Guofang Innovation, Honghui Fund, Huakong Fund, Hangzhou Fucheng, and others excessively increasing their investment.
This is a team of Tsinghua University teachers and students: Founder and CEO Ai Zhiyuan and CTO Chen Xianglin both graduated from the High Performance Institute of Tsinghua University's Department of Computer Science; Chinese Academy of Engineering Academician Zheng Weimin, also from the High Performance Institute, serves as the Chief Scientific Advisor; Tsinghua University Professor Wu Yongwei serves as the Chief Scientist; Tsinghua University Associate Professor Zhang Mingxing, as a co-founder, has been consistently guiding the company's technology strategy and key R&D breakthroughs, continuously driving cutting-edge technology advancements.
Three years ago, most domestic AI startups focused on large models, and even those in the AI Infra field mostly focused on training. Qujing Technology, however, chose to start from the large model inference track, building a high-quality AI Token factory. Now, as the demand for AI Tokens grows exponentially, this once-hidden track has finally caught fire. Similar to Zhipu, Qujing Technology has completed the equity investment transfer from Tsinghua University, becoming a typical project of Tsinghua's scientific and technological achievements transformation.
In just half a year, Qujing Technology has raised over 10 billion cumulatively. One after another, high-quality AI Token factories at the hundreds of billions and trillions level, built with Qujing's participation, have been completed, creating a new landscape for the AI industry.
Deep Integration of Production and Research, Building High-Quality AI Token Factories
Time goes back to 2023, when ChatGPT ignited the global generative AI wave. Witnessing this historic opportunity, Tsinghua University Professor Wu Yongwei from the Department of Computer Science and Ren Xuyang, founder of Zhenzhi Capital, decided to jointly initiate Qujing Technology. Their starting point in technology was precisely the Tsinghua University High Performance Computing Research Institute.
At the end of December the same year, Qujing Technology was formally established. The company's founder and CEO, Ai Zhiyuan, holds a Ph.D. from Tsinghua's High Performance Computing Institute. He previously served as R&D head in several key departments of a listed company, including big data, digitalization, and AI applications, accumulating comprehensive industry experience from technology R&D to large-scale implementation. Co-founder Zhang Mingxing, an associate professor at Tsinghua University, primarily conducts research in computer system architecture and was deeply involved in the foundational system construction of a leading large model company. As the company entered a phase of accelerated marketization, in March of this year, Dr. Wu Wenjie assumed the role of President of Qujing Technology. As a senior finance and strategy expert in the industry with a Ph.D. in Finance from the University of Hong Kong, she further strengthened the company's capabilities in financial management and global operations. Thus, a core team combining technical background, commercial perspective, and industry experience was formed.
While focusing on AI, the team made a choice that was considered non-mainstream at the time—while most AI entrepreneurs chose to invest in large model training, Qujing Technology focused on AI inference from the start. Simply put, training is about creating a "smart brain," while inference is about how to use that brain efficiently.
"Training is a cost item, inference is a profit item," Ai Zhiyuan explained. Their judgment at the time was that inference is what truly generates economic benefits and would be a broader market. What Qujing Technology aims to do is become the best partner for building and operating Token factories in the AI era, making the process of using the "brain" more efficient.
This is also Qujing Technology's positioning—compared to other AI Token factories, Qujing aims for the production of high-quality AI Tokens. Ai Zhiyuan further explains that when large models truly enter the production stage, what customers need is no longer a large model that "can chat," but one that can stably, efficiently, and cost-effectively complete real business tasks.
AI Tokens with real enterprise-level implementation value need to continuously meet multiple requirements such as low first-token latency, high concurrent load capacity, stable output quality, structured result generation, and function calls on models with parameters at the hundreds of billions or even trillions level, while also keeping the unit generation cost within an acceptable range for enterprises.
Any one of these capabilities alone is not the most difficult to achieve, which is the choice of most AI infra companies. But the real challenge lies in simultaneously meeting these metrics under real production loads and maintaining stability over long-term operation. According to data calculations, different combinations of capabilities can result in productivity gaps ranging from several times to dozens of times.
To achieve this goal, Qujing Technology employs globally pioneering technologies such as "full-system heterogeneous collaboration," "computing with storage," and "virtual-reality isomorphism" to build end-to-end capabilities covering heterogeneous integration, intelligent orchestration, and elastic scaling. Rather than focusing on a single pain point, they optimize every link in the AI Token production chain, ultimately achieving order-of-magnitude efficiency improvements.
Based on this, Qujing Technology also proposes the Token as a Service (TaaS) concept and, with its self-developed high-performance AI Token production service platform, the ATaaS platform, as the core, breaks through the conversion bottleneck between computing hardware investment and AI Token production capacity through system architecture and engineering capabilities, continuously and stably outputting high-quality AI Tokens like a standardized production line.
Over the past two years, this team has rarely appeared in the public eye. But Qujing Technology's choice is being validated by the industry—when Tokens become the currency of the AI era, the best timing has finally arrived.
Business Explodes, Over 10 Billion Raised in Half a Year
Investors began flocking to their doorstep.
Looking closely: in February this year, Qujing Technology completed its Angel++ round of financing invested by Parallel Technology; in May, it completed its Pre-A round, with an expanding lineup of investors including Xinglian Capital and Huakong Fund jointly leading the round, followed by Honghui Fund, Tianhao Energy, Shangshi Capital, Tianjin Ren'ai Hongsheng, Hangzhou Fucheng, and others. Existing shareholder Hillhouse Capital's venture arm, GL Ventures, continued to increase its investment.
In the latest round, investor enthusiasm remained high: led by Henan Investment Group's Huirong Fund, with old shareholders including Zhenzhi Capital, Shangshi Capital, Xinglian Capital, Shanghai Guofang Innovation, Honghui Fund, Huakong Fund, Hangzhou Fucheng, and others continuing their excessive additional investment. Thus far, within half a year, Qujing Technology has raised over 10 billion cumulatively. Investment Circle has learned that the company's next round of financing is already on the way.
In sight, more and more mainstream institutions are choosing to bet on Qujing, casting their votes for the future in advance. The continuous additional investments from existing shareholders serve as the strongest endorsement of Qujing's industry judgment, technical capabilities, and phased achievements.
Behind this, Qujing's initial judgment has become reality: with the rapid adoption of AI Coding, OpenClaw, etc., large-scale inference demands are accelerating, and AI commercial implementation is experiencing an overall explosion. The team's globally leading technological innovations in areas like computing with storage, full-system heterogeneous collaboration, and virtual-reality isomorphism are entering a window of value realization, ultimately reflected in the production efficiency of high-quality AI Tokens—the stronger the technical capability, the higher the inference efficiency, the lower the unit Token cost, and the greater the profit margin for enterprises.
Thus, Qujing Technology became sought after in the venture capital circle. Investors increasingly recognize the path the team has adhered to since its inception—"fewer models, deeper optimization." Qujing's focus is not on expanding the number of models, but on selecting key large models for in-depth refinement for real production scenarios, continuously improving their performance, stability, cost efficiency, orchestration capabilities, and cluster operation levels.
Ai Zhiyuan gives a vivid analogy: compared to building a "hypermarket" that sells everything, Qujing hopes to become more like a "boutique specialty store." Rather than constantly expanding the number of models, the company wants to concentrate resources on a few high-productivity models and high-value scenarios, enabling the same amount of computing power to continuously output more high-quality AI Tokens.
The underlying business judgment is that enterprise customers ultimately pay for business results, not for model compatibility numbers. In fact, the competitive landscape of large models is gradually converging. "Currently, less than 10% of the leading domestic models occupy the vast majority of the AI Token market." Based on this judgment, Qujing focuses its resources on a few leading models and core scenarios, thereby achieving the compound effect of continuous optimization.
Investment Circle obtained a set of data: since the Spring Festival of 2026, Qujing Technology's average AI Token production efficiency per unit of computing power has increased by more than 3 times, and the total output of high-quality AI Tokens has increased by more than 30 times. Among them, a certain leading trillion-parameter large model has already achieved a daily production capacity of trillions of high-quality AI Tokens. Meanwhile, the single-month revenue for June 2026 alone has already surpassed the entire year of 2025, and the revenue scale continues to grow at a high speed.
In Qujing Technology's view, what AI infrastructure ultimately competes on is not just who owns more GPUs, nor how many models are supported, but who can continuously produce more, more stable, and higher-quality AI Tokens. These are precisely the capabilities that investors collectively value.
Token is King, Welcoming the New AI World
AI inference is booming.
Data from the National Data Bureau shows that as of March 2026, China's daily Token call volume has exceeded 140 trillion, an increase of over a thousand times compared to two years ago. The core factor driving the explosion of AI Tokens is precisely the overall expansion of inference demand. This also confirms the judgment of investors: AI inference will become one of the largest global markets.
When Tokens become the "water, electricity, and coal" of the AI era, a hidden yet high-growth underlying business is emerging—the "AI Token factory." The logic is simple: in the future, those who can supply high-quality AI Tokens in a lower-cost, more stable, and more controllable manner will gain a first-mover advantage in the new Token economy track.
Thus, an AI Token factory "land grab" is unfolding across the country. As the track becomes increasingly crowded, commercialization becomes a question that AI infra enterprises must answer.
To some extent, Qujing has a greater ambition: not just to become an AI Token factory, but to be the designer, builder, producer, and operator of AI Token factories, becoming an indispensable player in the AI Token ecosystem.
This is also reflected in Qujing's two business models: one is the direct operation model, where after leasing or acquiring computing resources, they directly produce high-quality AI Tokens and supply them to leading model providers, internet platforms, AI application companies, and large enterprise customers, etc., obtaining higher returns by improving AI Token production capacity and operational efficiency; the other is the co-operation model, targeting clients planning to or already owning computing resources, undertaking the overall planning, system integration, construction delivery, and subsequent co-operation of AI Token factories.
Ai Zhiyuan further explains that more and more listed companies, state-owned enterprises, central enterprises, and various local intelligent computing centers hope to transition from traditional computing power leasing to higher value-added AI Token production. But there aren't many teams with real capabilities in inference system design, heterogeneous computing, and operations. What Qujing provides is a complete set of AI Token factory design and construction solutions, helping partners complete the transition from "selling computing power" to "selling high-quality AI Tokens."
A real AI Token economy should not be about one company completing all links alone, but about allowing more industry partners to participate and jointly build the ecosystem. It is evident that Qujing has already become a key link in this ecosystem, connecting model, computing power, application, and other industry partners, promoting the AI industry's shift from single-point innovation to ecological symbiosis.
This is also the inevitable direction of industry evolution. Every time a technological wave truly changes the world, it often doesn't happen at the moment of the new technology's birth, but rather when a set of supporting infrastructure matures, leading to a qualitative change. Just as the steam engine era needed railways; the internet era needed optical fibers and data centers. The AI era similarly needs a brand-new set of infrastructure to allow intelligence to flow stably, efficiently, and cost-effectively to all industries.
Behind this lies immense business opportunities embedded in the new AI infrastructure track. Only by deeply cultivating underlying innovation and collaborating within the industry ecosystem can one navigate technological cycles and continuously release long-term industrial and commercial value.
This article is from the WeChat public account "Investment Circle" (ID: pedaily2012), author: Wu Qiong.
