The arrival of the era of intelligent agents is opening an unprecedented strategic window for digital renminbi.

Written by: Wang Yang

Source: Ta Kung Pao

On March 24, 2026, Liu Liehong, Director of the National Data Bureau, revealed a shocking set of numbers at a press conference held by the State Council Information Office: at the beginning of 2024, the daily average token call volume in China was 100 billion; by the end of 2025, it surged to 100 trillion; and this March, it has already surpassed 140 trillion—within two years, a growth of over a thousand times.

Two years. A thousand times.

However, understanding this set of numbers merely through "growth" is far from sufficient. A deeper transformation is taking place, hidden within the dual meaning of the word "Token."

In the field of artificial intelligence, a Token is the basic unit of measurement for large models processing language—Token is the "atom" of the intelligent economy. In the field of digital finance, a Token is a digital certificate on the blockchain—token, is the "atom" of digital assets. These two Tokens, seemingly belonging to different technical lineages, are merging at an unexpected speed. When the production and consumption of AI tokens give rise to a colossal new economic sector, the settlement and assetization of this sector naturally require a digital currency infrastructure. Understanding this "dual Token integration" is the key to comprehending China's next phase of digital economy strategy.

The Chinese Weight of the Token Economic Revolution

The rise of Agentic AI (Intelligent Agent AI), marked by OpenClaw, is the explosive technological engine behind this transformation. Unlike the traditional linear model of question-and-answer, intelligent agents enter a continuous working loop: reading data, breaking down tasks, writing code, calling tools, detecting errors, reflecting and adjusting, executing again—each step consumes Tokens. A single intelligent agent session for a complex software project may consume dozens or even hundreds of times more Tokens than a regular conversation. IDC predicts that by 2030, there will be 2.216 billion active AI intelligent agents globally.

The open-source nature of OpenClaw makes it naturally compatible with China's AI ecosystem—it can seamlessly integrate with domestic large models like DeepSeek and connect to local applications like WeChat through customized configurations. Alibaba, Tencent, and ByteDance quickly integrated it into their own cloud services and application ecosystems. This combination of "open-source framework + low-cost domestic models" has rapidly made China the largest market for AI Token consumption in the world.

This explosive demand has been confirmed by international data. Data from the OpenRouter platform shows that in February 2026, the call volume of AI models in China surpassed that of the United States for the first time—during the week of February 9 to 15, the call volume in China reached 41.2 trillion Tokens, exceeding the 29.4 trillion Tokens of the U.S. during the same period. Subsequently, Chinese models maintained their lead, with 46.9 trillion Tokens in the week of March 15, ranking first in the world. Models such as MiniMax, GLM, and Tongyi Qianwen have surged to the top of the Token usage rankings, essentially a result of price advantages combined with an explosion in intelligent agents.

The technological backdrop of this reversal is the cluster explosion of China's open-source large models. According to media reports citing research data from RAND Corporation, after the release of DeepSeek R1, China's global market share of AI model access skyrocketed from 3% to 13% within two months. JPMorgan Chase predicts that the annual compound growth rate of China's AI inference Token consumption will reach 330% between 2025 and 2030, representing an increase of approximately 370 times over five years.

An Emerging Value-Added Chain

The underlying supply capacity of Tokens in China is built upon a complete industrial value chain from energy to intelligence: electricity—computing power—large models—Tokens—intelligent applications. Each link is experiencing real value-added growth.

The electricity layer is at the top of the chain. By the end of 2025, China's cumulative installed power generation capacity is expected to reach 3.89 billion kilowatts, a year-on-year increase of 16.1%; the total electricity consumption in society has exceeded 10 trillion kilowatt-hours for the first time; and China's electricity prices are highly competitive globally. The surge in electricity demand from AI data centers has elevated this tier from basic infrastructure to a strategic asset—IDC predicts that by 2027, the IT energy consumption of China's AI data centers will rise to 146.2 TWh.

The computing power layer is the critical "refining" segment. By the end of 2024, China's total computing power is expected to reach 280 EFLOPS, ranking second globally, with intelligent computing power accounting for over 30%. In 2026, "collaboration between computing and electricity" will be explicitly included in the government work report, with top-level design placing computing power and electricity on par as strategic infrastructure.

The model layer serves as the core factory for generating high-value-added Tokens. A cluster of open-source and closed-source models such as DeepSeek, Tongyi Qianwen (Qwen), Wenxin Yiyan, Hunyuan, and Kimi forms a rare high density of the large model industry ecosystem globally.

The Token layer is the end product of this chain, yet is distinctly different from oil: it has no physical form, can be infinitely replicated, can be transmitted across borders instantly, and the marginal production cost approaches zero. This means that the high value-added of the entire value chain is extremely concentrated in the model layer and application layer—whoever masters the highest quality models holds the pricing power of Tokens.

The National Data Bureau disclosed that by the end of 2025, over 100,000 high-quality datasets have been built nationwide, totaling more than 890PB, approximately 310 times the total amount of digital resources in the National Library of China. These datasets serve as the "raw material reserves" for the continuous iteration of large models, constituting the invisible strategic depth of the entire value chain.

Intelligent Agents Begin Trading

Value chain analysis reveals the production logic of Tokens, but a more disruptive trend is emerging at the end of the value chain: intelligent agents are transforming from consumers of Tokens to autonomous participants in economic activities.

Economic activities between machines have begun to sprout. Moltbook—a social network designed specifically for AI intelligent agents—has witnessed remarkable phenomena: intelligent agents autonomously publish content, comment, vote, and even issue cryptocurrency tokens independently. Former Tesla AI director Andrej Karpathy stated that the activities on Moltbook represent "the most sci-fi-like thing" he has seen recently.

When an intelligent agent autonomously completes shopping, pays service fees, and calls APIs of other intelligent agents for users, it requires a settlement layer. This settlement layer must support three characteristics: micropayments (the cost of a single API call may be less than a cent), programmable conditions (settlement occurs only if the task is successful), and real-time clearing (intelligent agents cannot wait two to three business days for bank clearing cycles). The traditional banking system cannot meet these needs, but stablecoins can. However, if intelligent agents default to using stablecoins such as USDT or USDC based on the dollar for settlement, and if global developers are accustomed to trading in dollar-valued API markets, then even if China's industrial chain value from electricity to Tokens to intelligent applications is operating, it will still function under the dollar system.

This is precisely where the strategic urgency of the "dual Token integration" lies: the AI Token economy requires a digital currency token layer to complete settlement and assetization—this layer must either be digital renminbi or dollar-pegged stablecoins. There are virtually no middle-ground options.

Digital Assets as a Stabilizing Force

At every node of this value chain, there are underlying assets that can be capitalized, securitized, and even digitized.

Power plants and power grids: The strong demand for electricity from AI data centers makes electricity itself a candidate for financial assetization and can be converted into digital assets in both the mainland and Hong Kong markets.

Computing power centers: RWA (Real World Assets, tokenization of real-world assets) focusing on GPU power leasing has been accelerating exploration.

Internationally, decentralized computing markets like Akash and io.net have tokenized GPU hours; Chinese background computing operators are also exploring similar pathways through compliant channels in Hong Kong.

Such AI value chain digital assets structurally possess three core characteristics: first, continuous cash flow—electricity, computing power, and Token consumption all represent predictable recurring revenue, forming the basis for asset pricing. Second, a strong anchoring to the real economy—the demand for Tokens arises from genuine intelligent agent tasks, underpinned by enterprises' real need for cost reduction and efficiency increase; the supporting computing power and electricity are "tangible" physical assets. Third, they are sizable and continuously growing—if just 1% of computing power assets realize financialization, it could correspond to thousands of billions of dollars in new asset categories.

The value chain from electricity to Tokens is poised to become the next generation of "government bond-level" stabilizing assets in the global digital asset market—just as government bonds serve as benchmark assets for pricing and collateral in the traditional financial system, AI value chain assets have the potential to become benchmark anchors in the digital asset market. What underpins it is genuine industrial production, rather than mere financial narratives.

The Strategic Opportunity for Digital Renminbi

On January 1, 2026, digital renminbi (e-CNY) completed a version leap from 1.0 to 2.0. The core of this upgrade is not only a technological iteration but a fundamental transformation in monetary attributes: upgrading from "digital cash" to "digital deposit money," with digital renminbi balances in commercial bank wallets officially starting to accrue interest.

This upgrade's timing coincides closely with the acceleration of the Token economy. In the Token economy, China simultaneously covers both the supply and demand sides—holding the dual position of "both a major seller and a major buyer," coupled with the global technological influence brought by open-source models, provides a realistic foundation for China to strive for a dominant position in the Token price formation mechanism.

If the domestic and foreign trading of Tokens, the issuance and circulation of computing power RWA, and the settlement of green electricity—computing power futures are all priced in digital renminbi, then digital renminbi will naturally be embedded in every segment of this value chain. The specific pathways include at least three:

First, real-time settlement of Token spot transactions. The clear Token consumption generated by AI model API calls and corresponding costs perfectly match the programmability and smart contract capabilities of digital renminbi, enabling "Token billing and real-time settlement," significantly reducing settlement friction and capital occupation of existing cloud computing bills. This is particularly crucial in the intelligent agent economy—an autonomously operating intelligent agent may initiate thousands of micro-transactions daily, each amounting to mere fractions of a cent; the traditional payment system simply cannot accommodate such granular settlement demands.

Second, cross-border RWA transaction settlements. Business entities along the Token value chain have the opportunity to issue and circulate RWA in compliance within the Hong Kong market, priced and settled in digital renminbi. This pathway aligns with the trend of tokenizing Tokens abroad while providing substantial asset anchors for the internationalization of digital renminbi. By the end of November 2025, the mBridge project had cumulatively processed 4,047 cross-border payments, amounting to approximately 387.2 billion yuan, with about 95.3% of the amount using digital renminbi—this precedent indicates that the technological feasibility has been preliminarily verified.

Third, the transformation of domestic digital asset infrastructure. Implementing the digitalization of the Token value chain assets through the digital renminbi blockchain and digital asset platforms naturally settles transactions in digital renminbi, improving China's financial asset structure, and enhancing transaction efficiency and regulatory precision.

Of course, this prospect is not without resistance. The compliance framework for computing power RWA is yet to be完善; domestic regulatory attitudes towards tokenized assets remain cautious; the technological compatibility of digital renminbi with blockchain-native assets faces substantial engineering challenges; and whether market participants can strike a balance between regulatory rhythm and technological implementation remains uncertain.

Global Competition: The Window is Open, but Not Forever

China's unique advantage lies in simultaneously possessing three valuable cards: a mature central bank digital currency infrastructure (digital renminbi 2.0), a rapidly growing ecosystem of AI models and agents (the largest Token consumption globally), and a vast application market. More critically, China has the institutional capability to systematically connect these three elements. This "trinity" condition is something that any other economy will find difficult to replicate in the short term.

However, the window of opportunity is limited. The settlement habits of the intelligent agent economy are forming. If global developers default to trading in the dollar-priced API market, and if the autonomous payments between intelligent agents typically flow towards dollar-pegged stablecoins, then even if China is leading in technology and scale, the economic value will still be captured by the dollar system. Digital renminbi must embed itself in the infrastructure layer of the intelligent agent economy at an early stage rather than wait until the landscape is solidified before attempting to replace it.

First, swiftly initiate a "digital renminbi smart contract sandbox" in the Greater Bay Area to pioneer trials for intelligent agent micropayment settlements, testing integration solutions between digital renminbi and AI intelligent agents in a controlled environment, accumulating practical experience and regulatory data.

Second, establish national standards for intelligent agent financial security, clarifying the authorization mechanism for autonomous payments, limit management, anomaly detection, and triggers for human intervention, providing a safety baseline for the intelligent agent economy.

Third, promote native API integration between digital renminbi and mainstream AI infrastructures, allowing the billing and settlement of domestic large model Tokens to directly incorporate digital renminbi payments, lowering the access threshold for developers, and making digital renminbi the "default currency" of the Token economy.

Fourth, within the framework of the "Belt and Road" digital economic cooperation, prioritize incorporating intelligent agent economy settlements into cross-border pilot scenarios for digital renminbi, utilizing the high penetration rate of Chinese open-source models in developing countries to promote digital renminbi as one of the main settlement currencies in the global AI economy.

The dual meanings of the word "Token" are not mere coincidences of language but are the inevitable convergence points of technological evolution. The intelligent agent economy driven by AI Tokens is experiencing explosive growth, requiring a matching programmable settlement infrastructure. From electricity to computing power to large models to Tokens to intelligent agents, this complete value chain in China is calling for a digital currency that coexists with it.

The arrival of the era of intelligent agents is opening an unprecedented strategic window for digital renminbi: to establish renminbi's pricing power and settlement position in a brand-new economic field where China has significant comparative advantages. The issue is not whether this "dual Token integration" will happen, but whether China can actively shape its rules.

The window has opened, but it will not remain open forever.