Introduction: Transforming domestic computing power resources into services, allowing overseas customers to invoke them via the network, and charging based on usage.
On the evening of October 18, an AI order worth USD 200 million (equivalent to CNY 1.338 billion) put Runjian Shares in the spotlight of Token export.
This A-share company, which primarily engages in communication network management and operation & maintenance and has been continuously expanding its computing power business in recent years, has just secured a cross-border AI service contract worth USD 200 million. The contract has a one-year term, during which overseas customers will invoke large model inference services via API interfaces, and all related computing power resources will be deployed in intelligent computing centers within China.
Notably, this contract includes a minimum purchase guarantee clause: if the service fee corresponding to the actual Token consumption by the customer within the contract year is less than USD 200 million, the difference must be made up as agreed.
The amount of this single order is equivalent to approximately 35.6% of Runjian Shares's operating revenue in the first half of 2026. What it sells is neither servers nor chips; what the customer purchases are the Tokens consumed during the operation of the large model.
In the global expansion of China's AI industry, a new way of making money is emerging: transforming domestic computing power resources into services, allowing overseas customers to invoke them via the network, and charging based on usage.
This transaction has also brought a concept that originally appeared more in AI industry discussions into the commercial contracts of listed companies.
In the past, market attention on the global expansion of China's AI often focused on whether large models could catch up with overseas competitors and whether domestic chips could enter the international market. Today, model invocation services, inference computing power, and Token billing are also beginning to become new directions for enterprises to explore overseas revenue.
For China's AI industry chain, this means new market space. For Runjian Shares, a large contract also brings a more practical test: when the intelligent computing centers built domestically begin to serve overseas customers, how much money can this business actually make?
After all, USD 200 million is just the contract amount. What truly determines the value of this transaction also includes computing power costs, equipment depreciation, customer payments, and whether the service provider can retain sufficient profit in the fierce price competition.
The story of Token export has already seen its first large order. Next, what the market needs to see is how this business can translate from contract amounts into real cash revenue and profit.
China's AI Global Expansion Has a New Path
To understand Token export, one must first clarify what exactly the customer is purchasing.
When users ask questions, generate text, or write code using a large model, the model processes the input and generates the output. Input and output are usually measured in Tokens, and the service provider then charges fees based on the model type, the quantity of input and output, and specific billing rules.
Developers can invoke the model via APIs and integrate it into their own applications. Overseas enterprises can therefore directly use AI capabilities without the need to build a complete model inference system themselves.
This model reduces the initial investment for customers to use AI and also provides model service providers with the opportunity to charge based on usage.
The increased activity of Chinese large models in overseas developer communities has provided a market foundation for this business.
Chinese AI companies such as DeepSeek, Yue Zhi An Mian, Zhi Pu, and MiniMax are expanding into overseas markets through different model products and technical routes. Model performance, invocation price, open-weight strategies, and developer tools will all influence the choices of overseas customers.
For developers, selecting a model often requires considering answer quality, response speed, invocation cost, and stability simultaneously. As long as the model can meet specific business needs and the price is competitive, it has the opportunity to enter application development and enterprise workflows.
However, there is still a distance between a model gaining attention and achieving commercial success. Developer trials, actual payments, enterprise procurement, and long-term renewals represent market progress at different stages. An increase in invocation volume also needs to be combined with price changes and cost structures to judge the quality of revenue growth.
Runjian Shares's current order provides another perspective for observation.
The company has long been engaged in communication network management and operation & maintenance, and has gradually expanded its digital network, energy network, and computing power network businesses. With the accumulation of capabilities in the construction and operation of intelligent computing centers, the company has begun to explore new computing power services such as Tokens.
In this contract, Wuxiang Yungu provides large model inference invocation services to overseas customers via APIs, while the computing power facilities remain in China. Customers can use the related resources remotely and settle according to the contract terms.
This business arrangement provides a feasible path for China's computing power infrastructure to participate in the global AI market. Even if an enterprise has not independently developed a foundational model, it may still participate by leveraging its computing power resources, technical operations, and service delivery capabilities.
Competing on Price, Computing Power, and Customer Payments
On the Token industry chain, at least two types of enterprises are looking for business opportunities.
The first type is model vendors. They provide API services for their self-developed models to customers, and their revenue is related to customer invocation volume, product pricing, and contract arrangements. The capabilities of the model, inference efficiency, and developer ecosystem determine whether the enterprise can continuously attract customers.
The second type is computing power service providers. They provide services centered around computing resources, inference deployment, API access, and operational delivery. Some enterprises own their intelligent computing centers, while others provide services by integrating computing resources from different sources.
Runjian Shares's current contract is closer to the latter model.
The capability requirements for the two types of enterprises differ. Model vendors need to steadily improve model performance and control the cost of each inference; computing power service providers need to do a good job in resource scheduling, equipment maintenance, and customer service to achieve higher utilization rates for existing computing resources.
The two can also form cooperative relationships. Models, computing power, interface services, and overseas customer channels can be provided by different enterprises respectively and then connected through commercial contracts.
This division of labor expands the space for China's AI industry chain to participate in overseas markets and also brings new competition.
When more and more enterprises can provide similar inference services, customers will compare invocation prices, response speeds, service stability, and technical support capabilities. If a service provider lacks obvious cost advantages or customer stickiness, its revenue growth is easily affected by price competition.
Therefore, Token invocation volume is just one indicator to measure business scale. Customer retention rate, single inference cost, service gross margin, and customer payment status equally determine the value of this business.
How Much Profit Can It Bring to Runjian Shares?
The scale of Runjian Shares's contract is quite considerable.
The company achieved an operating revenue of CNY 3.756 billion in the first half of 2026, and the USD 200 million contract amount is equivalent to approximately 35.6% of its operating revenue in the first half of the year.
However, there is an obvious difference between the contract amount and current revenue and profit.
According to the announcement, the contract term is from October 19, 2026, to October 18, 2027. The service adopts a post-paid model based on usage, with natural months as the settlement cycle. The final confirmed revenue amount needs to be determined based on actual performance and accounting treatment.
The contract also sets an annual minimum purchase guarantee clause. If the service fee corresponding to the actual Token consumption by the customer within the contract year is less than USD 200 million, the difference must be made up as agreed. This provides a certain procurement guarantee for the contract, but customer payments, contract fulfillment, and revenue recognition still require subsequent verification.
The profitability performance of Runjian Shares's existing computing power business can also help understand the cost pressures faced by this order.
According to the company's 2026 semi-annual report, the computing power network business achieved an operating revenue of approximately CNY 519 million in the first half of the year, a year-on-year increase of 50.31%; the operating cost was approximately CNY 483 million, a year-on-year increase of 55.08%; and the gross margin was 6.97%, a year-on-year decrease of 2.86 percentage points.
Revenue grew rapidly, but the cost growth rate was even faster. The company stated that the low gross margin of the Token factory business is mainly affected by factors such as high depreciation during the fixed asset commissioning period.
This is related to the operational characteristics of computing power services. Server procurement requires a large amount of capital, and equipment continuously generates depreciation after being put into use. Electricity, operation & maintenance, and resource scheduling also affect costs. Even if customer demand grows, enterprises still need to improve equipment utilization rates to enhance profitability.
A simple scenario calculation can be made using the existing business gross margin: assuming the USD 200 million contract amount ultimately forms revenue in full, and the gross margin reaches the 6.97% of the computing power network business in the first half of the year, the corresponding gross profit would be approximately CNY 93.2 million.
Of course, the above is just a simple scenario calculation. The commercial value of this order ultimately depends on whether the company can deliver the service at a reasonable cost and retain sufficient gross profit from the revenue.
Having Computing Power Alone Is Far from Enough
Token services may seem to be deliverable quickly via APIs, but behind the scenes, a complete set of operational capabilities is still required.
First is the computing power supply. Large model inference requires continuous response to customer requests. Equipment failures, insufficient resources, or service interruptions can all affect the customer experience. Service providers must ensure the stable operation of computing resources, networks, and systems.
Second is customer expansion. A single contract can generate phased revenue, but only long-term renewals and continuous procurement have the opportunity to establish a stable business scale. Service providers also need to prove that they possess sufficient price competitiveness and service capabilities to make customers willing to continue using their services.
Cross-border business will also bring additional requirements. Service providers need to pay attention to customer identity, service content, data processing methods, and regulatory requirements in relevant markets. The deployment of computing power domestically does not mean that cross-border services can ignore compliance and operational risks.
Runjian Shares also pointed out in the announcement that the Token business places high demands on capital investment, technical capabilities, and customer expansion capabilities. Market competition is fierce, and there is still uncertainty regarding future order acquisition and large-scale commercialization.
The company's financial status further illustrates the capital investment required for the transformation.
In the first half of 2026, the net cash flow generated from operating activities of Runjian Shares was approximately CNY -2.876 billion. The company explained that the expansion of the net cash outflow is mainly related to increased investments in new businesses such as computing power and energy, as well as the concentration of customer payments in the second half of the year.
This set of data needs to be understood in the context of the company's business characteristics, but it also reminds the market: computing power service providers often need to invest capital first to build or prepare resources, and then gradually recover costs through subsequent services.
If customer demand continues to grow and existing equipment is fully utilized, the depreciation of fixed assets has the opportunity to be apportioned across more business. If customer procurement falls short of expectations and equipment utilization rates are low, depreciation and other fixed costs may drag down profits.
Therefore, service providers need to do two things simultaneously: secure stable customer demand and control the delivery cost per unit of Token.
Can It Become a New Export for China's AI?
Token export provides a new internationalization method for China's AI industry.
Model vendors can sell model capabilities to overseas customers via APIs, while computing power service providers can rely on domestic intelligent computing centers to provide inference services. As application development and enterprise deployment increase, a more detailed division of labor may form among models, computing power, and service delivery.
This is especially important for enterprises that have already invested heavily in building intelligent computing infrastructure. Overseas customers can increase potential sources of demand, help enterprises expand their service scope, and also find more commercial applications for existing computing resources.
However, whether this business can achieve sustained growth still depends on several practical indicators.
First, whether overseas customers can continue to procure and gradually form renewals and repurchases; second, whether the gross margin of the computing power network business can improve, proving that revenue growth is beginning to bring better profitability; third, whether contract revenue can be transformed into stable cash flow to cover computing power investments and daily operational costs.
These three indicators reflect the maturity of the business model better than a single contract amount.
Runjian Shares's USD 200 million order indicates that Token cross-border services have entered the stage of specific commercial contracts.
Next, the company needs to use actual fulfillment, customer payments, and profitability performance to prove that this order can form sustainable business results.
For the entire industry, opportunities are also accompanied by competition. Model technology continues to iterate, inference prices may change, and overseas customers will continuously compare the costs and quality of different service providers. Only by establishing advantages in service capabilities, cost control, and customer relationships can enterprises gradually accumulate long-term value from short-term orders.
Token export is becoming a new business. Whether it can develop into a stable source of overseas revenue for China's AI industry ultimately depends on whether enterprises can transform computing resources into continuously paid services and maintain reasonable profits and cash flow during the expansion process.