"Token Economy" Rises as Listed Companies Increase Full-Chain Layout

When it comes to stock trading, rely on Golden Kylin Analysts’ reports—authoritative, professional, timely, comprehensive—helping you uncover potential thematic opportunities!

The concept of “tokens” continues to ferment, with the A-share computing power sector gaining strength. On March 25, Shenzhen Tefa Information (rights protection) Co., Ltd., 263 Network Communications Co., Ltd., Shanghai Tianji Technology Co., Ltd., and other stocks hit the daily limit. Beijing Guanghuan Xinwang Technology Co., Ltd., Guangdong Aofei Data Technology Co., Ltd., and several others saw significant gains.

Tokens are the smallest units of information processing in large models, possessing measurable, tradable, and quantifiable attributes. They are the fundamental units for AI service billing and settlement. If a large model is an intelligent machine, tokens are like the “kilowatt-hours” it consumes during operation and the basic measurement unit for billing when providing services externally.

Data from the National Bureau of Statistics shows that the daily token invocation volume in China has experienced explosive growth, reaching an average of 100 billion per day in early 2024, soaring to 100 trillion by the end of 2025, and surpassing 140 trillion by March 2026—more than a thousandfold increase in two years.

The “Token Economy” is expected to reshape AI’s commercial logic, shifting large model companies from “selling capabilities” to “selling usage.” Yuan Shuai, Deputy Secretary-General of the Zhongguancun Internet of Things Industry Alliance, told Securities Daily that under this model, company revenue no longer depends on long-term project delivery but on real-time, concurrent interaction scales. This shift can drive the AI industry from purely technology-driven to finely tuned operational management, enabling high-frequency, high-engagement applications to rapidly transform into explosive performance growth.

In response to the potential business revolution brought by the “Token Economy,” global tech giants are accelerating strategic deployment. Nvidia founder and CEO Jensen Huang proposed “Token Economics” at the 2026 GTC conference, believing that future data centers will no longer just store files but will be “factories” producing tokens. Nvidia also released a full-stack hardware system to reduce token production costs.

On March 16, Alibaba Group Holding Limited officially established the AliBABATokenHub (ATH) business group, focusing on “creating tokens, delivering tokens, and applying tokens,” directly led by Group CEO Daniel Zhang.

“The Token Economy will bring huge development opportunities,” said Li Peng, Senior Vice President of Huawei Technologies Co., Ltd. He predicts that by 2030, China’s AI-related industry scale will surpass 10 trillion yuan.

Domestic listed companies are accelerating their full industry chain deployment around token production, invocation, distribution, and settlement—focusing on upstream computing infrastructure, midstream platform services, and downstream applications. The industry concentration continues to rise.

Upstream computing infrastructure is the core foundation of token production and the most directly benefiting segment. For example, Xiechuang Data Technology Co., Ltd. (hereafter “Xiechuang Data”) achieved significant year-over-year growth in revenue and net profit in 2025. Last year, the company planned to invest 21.2 billion yuan in high-performance servers, massively expanding its computing clusters.

Geng Kangming, Chairman of Xiechuang Data, told Securities Daily that current AI industry giants are optimistic about sustained growth in computing power. Companies like OpenAI, Anthropic, Microsoft, and Google are deploying intelligent agents, with reasoning models like DeepSeek pushing agents into practical deployment.

The midstream platform acts as a scheduling and settlement hub, responsible for token distribution, billing, and computing resource scheduling. Take Zhongbei Communication Group Co., Ltd. as an example; it has scaled operations, currently managing over 17,000 PetaFLOPS of computing power, providing intelligent computing services to clients like Kingsoft Cloud Holdings.

In downstream application fields, listed companies leverage industry advantages to integrate token technology with vertical scenarios, accelerating commercialization. Major AI application providers such as iFlytek rely on large models like Xinghuo to achieve large-scale monetization in education, healthcare, and office scenarios. Shenzhen UBTECH Robotics Co., Ltd. uses large models to empower robots, supporting intelligent interaction and embodied intelligence deployment.