Artificial intelligence tokens — once a term familiar only to techies — are rapidly entering everyday life in China.
Banks, cafes, and telcos, among other small and big businesses, are offering free AI tokens as incentives to attract more customers. Some businesses are selling tokens much like mobile data plans, offering them as credit card rewards, or giving them out like free Wi-Fi.
An AI token is essentially a unit of computing that an AI model uses to process text. When a user asks a model a question or gives it an instruction, the words are broken down into small pieces called tokens, which the model processes to generate a response.
500 trillion Daily token consumption in China in mid-2026.
The massive popularity of AI has led to has sharply increase in token consumption globally. Daily token consumption in China surged to 500 trillion in mid-2026 from 100 billion in early 2024.
Token prices have become a new front in the global AI race as U.S. and Chinese AI companies aggressively lower them to compete for wider adoption of their models. Chinese open-source and open-weight models, despite lagging behind U.S. rivals on some measures of frontier performance, can cost 60% to 90% less than comparable models from OpenAI and Anthropic. Cheaper tokens allow Chinese businesses to experiment with ways to package and market them to ordinary consumers.
Poe Zhao, a Beijing-based analyst and founder of the newsletter Hello China Tech, told Rest of World some of the token packages Chinese companies pushed into the consumer market are a “supply-led experiment.”
“Most ordinary users still encounter AI through an app or feature, without seeing a token balance,” he said.
Zhao said China’s effort to bring tokens into the consumer marketplace points to an unusual approach to commercializing AI.
Here are five ways AI tokens are moving into everyday life in China:
AI tokens as credit-card rewards
In July, Chinese AI startup Moonshot AI launched a credit card named after its model Kimi, in partnership with the Agricultural Bank of China and American Express. Users earn Kimi AI tokens, as well as agent and coding quotas, on their spending, just like credit cards reward consumers with airline miles and cashback offers.
The trend began in June, when China Merchants Bank launched a credit card targeting AI developers. It offered new cardholders up to 1.8 billion tokens for use on Chinese AI startup MiniMax’s models.
Shanghai Pudong Development Bank also launched a credit card for AI developers, offering subsidies worth up to 3 billion tokens for AlibabaAlibabaAlibaba, founded in 1999 by Chinese entrepreneur Jack Ma, is one of the most prominent global e-commerce companies that operates platforms like AliExpress, Taobao, and Tmall.READ MORE’s Qwen models.
AI tokens as mobile data
China’s three major telecom operators have packaged AI usage much like mobile data plans.
China Telecom offers consumer plans starting at 9.9 yuan ($1.40) a month for 10 million tokens, with larger allowances available for heavier users and businesses. The company also has a TokenHub platform that offers subscriptions on 142 large models, as the state-owned carrier tries to evolve from selling traditional telecom connectivity to AI computing and services.
Its competitors offer similar plans. China Mobile offered 400,000 tokens for 1 yuan (14 cents) in Shanghai, payable through users’ phone bills. China Unicom introduced monthly plans ranging from 15 yuan ($2.10) for 6 million tokens to 45 yuan ($6.30) for 18 million for business users.
AI tokens as a freebie
Instead of offering free Wi-Fi, an increasing number of Chinese restaurants, cafes, and bars are now luring customers with free AI computing.
At Beijing’s AGI Bar, customers who buy a drink and connect to the Wi-Fi can get unlimited access to DeepSeek V4 Flash. The bar, which targets customers who are programmers and AI startup founders, runs the model locally using a Nvidia DGX Spark workstation.
Nearby, Beijing’s Jingu Yuan dumpling restaurant gives customers 10 yuan ($1.40) worth of computing credits after a meal. Cards on its tables tell diners: “You’re full after eating the dumplings. Go collect some tokens at the cashier so your [AI] agent can be fed with some computing power too.” Its two stores have been handing out more than 100 vouchers a day.
In Changsha, an upcoming 24-hour coffee shop uses a similar approach for promotion, offering customers AI tokens with their coffee.
AI tokens on e-commerce
Tokens have also reached one of China’s largest online secondhand marketplaces. On Alibaba-owned Xianyu, a platform known for selling used phones, furniture, and clothes, sellers advertise cheap access to AI models in packages measured by tokens.
Listings offer millions of tokens for a few dollars in day passes or monthly subscriptions and shared access to AI accounts. Some sellers break up model allowances and resell them to students, independent developers, and small businesses. Others attempt to profit from the gap between official token prices and the discounted or free access they obtained elsewhere.
AI tokens in bank lending
The increasing number of AI startups in China has presented a challenge for its banks, as they struggle to assess the financial status of young companies on traditional metrics such as physical assets.
Banks in the southern city of Guangzhou are experimenting with tokens to boost business for AI startups. The city’s Haizhu district launched a financial product called “token loan” in August. It allows banks including Bank of China, China CITIC Bank, and Bank of Guangzhou to measure a company’s token production and consumption to decide how much it can borrow.
Facts Only
* AI tokens are a unit of computing used by an AI model to process text.
* Daily token consumption in China is projected to reach 500 trillion in mid-2026, up from 100 billion in early 2024.
* Chinese open-source and open-weight models can cost 60% to 90% less than comparable models from OpenAI and Anthropic.
* Moonshot AI launched a credit card named after its model Kimi, partnered with the Agricultural Bank of China and American Express.
* China Merchants Bank launched a credit card targeting AI developers offering up to 1.8 billion tokens for use on MiniMax models.
* Shanghai Pudong Development Bank offered subsidies worth up to 3 billion tokens for AlibabaQwen models.
* China Telecom offers consumer plans starting at 9.9 yuan ($1.40) per month for 10 million tokens.
* China Mobile offered 400,000 tokens for 1 yuan in Shanghai.
* China Unicom introduced monthly plans ranging from 15 yuan ($2.10) for 6 million tokens to 45 yuan ($6.30) for 18 million for business users.
* Alibaba-owned Xianyu allows sellers to advertise access to AI models in packages measured by tokens.
* Banks are experimenting with "token loans" in the Haizhu district of Guangzhou to measure token production and consumption for assessing startup financial status.
Executive Summary
The proliferation of AI tokens is entering the Chinese consumer market through various commercial incentives offered by businesses. These tokens function as a unit of computing used by AI models to process text, and their value is being leveraged to attract customers. Businesses are employing diverse monetization strategies, such as offering tokens as credit card rewards, bundling them with mobile data plans, providing free access via services like free Wi-Fi, and incorporating them into e-commerce listings and banking products.
Token consumption in China is projected to reach 500 trillion daily tokens by mid-2026, reflecting the rapid growth in AI adoption. This trend is fueled by the ability of Chinese companies to offer access to open-source models at significantly lower costs compared to Western competitors. Analysts suggest that these token packages represent a supply-led experiment in commercializing AI, moving the experience toward ordinary consumers who interact with AI primarily through applications rather than directly managing token balances.
The integration of tokens spans consumer services, telecommunications, retail, and financial lending, demonstrating an attempt to embed AI infrastructure into existing economic and service frameworks across China.
Full Take
The movement toward commercializing AI via tokens reveals a fundamental tension between the underlying technology's complexity and its packaging for mass consumption. The shift from technical metrics to token-based rewards suggests an attempt by entities to democratize access—making abstract computational resources tangible, quantifiable, and exchangeable like currency or data plans. This pattern appears driven by competitive dynamics: Chinese developers leverage cost advantages in open-source models to create scalable economic incentives that Western counterparts cannot easily match.
The various application methods—credit rewards, bundled data plans, freebies, e-commerce access, and lending metrics—demonstrate a strategy of embedding AI access into existing, deeply ingrained consumer touchpoints. This is not merely an innovation in AI; it is an infrastructural negotiation where large technology entities, state-owned telecom providers, and traditional financial institutions are adopting new mechanisms to capture value in the emerging AI ecosystem. The focus on "supply-led experiments" suggests that the mechanism of distribution (tokens) is as strategically important as the underlying model performance itself.
The deeper implication for human agency lies in whether these token systems truly empower users or merely create new layers of dependency and quantification over access to essential computational power. If AI access becomes entirely mediated by transactional tokens, the locus of control shifts from utilizing the intelligence to managing one's digital ledger and credit against that intelligence. The pattern suggests a consolidation where abstract intellectual property is rapidly being translated into fungible, state-managed, or institutionally controlled commodities to drive adoption.
What steps are needed to ensure that these token systems serve equitable access rather than creating new forms of digital stratification? How do the established financial and communication structures adapt to accommodate assets defined purely by computational utility? What mechanisms can be developed to ensure that the pursuit of lower costs does not inadvertently prioritize systemic control over genuine user autonomy in the AI age?
Sentinel — Human
The text presents a detailed, structured analysis of the emerging concept of AI tokens in China by synthesizing specific business and operational examples with large-scale consumption figures.