By 2026, the most profound transformation in the AI industry will not lie in the technology itself, but in the unit of measurement. According to disclosures from the National Data Bureau, China's daily Token call volume exceeded 140 trillion in March 2026, a more than 1000-fold increase from 100 billion at the beginning of 2024. Tokens are replacing traffic as the industrial currency of the AI era. However, the value between different Tokens varies significantly. Computing Tokens and Model Tokens are trending towards thin margins, while Scenario Tokens—the value unit closest to business outcomes—represent the true high-margin ground for the AI application layer. MARKETINGFORCE's (02556) strategy is clearly oriented towards this direction.
The three-layer structure of Tokens is: Computing Token → Model Token → Scenario Token. Token economics is reshaping value distribution in the AI industry chain, but the pricing logic differs entirely across layers. The Computing Token, the unit of GPU computing power consumed to run AI models, is the most fundamental energy source of the AI industry. This layer is highly standardized and homogeneous. With the large-scale construction of computing infrastructure and technological advancement, computing costs are on a long-term downward trend, ultimately leading to a low-margin structure akin to utilities. The Model Token is the billing unit for large model API calls. While more differentiated than Computing Tokens, the price point for Model Tokens is also decreasing due to the open-sourcing and homogenization of model capabilities, alongside the proliferation of multi-model fusion scheduling frameworks. The Scenario Token, however, is the value unit that translates underlying computing power and model calls into quantifiable business outcomes within specific operational contexts. For enterprise clients, the basis for continued payment is not "how many Tokens were called," but whether those Tokens lead to higher conversion rates, faster deal progression, lower customer service costs, and improved organizational efficiency.
Performance validation indicates that MARKETINGFORCE's Scenario Token capability is already materializing. The company's Q1 business update disclosed on April 15th shows that for Q1 2026, AI application business revenue increased approximately 110.5% year-over-year, while gross income from precision marketing services grew about 0.9%. It is evident that the company's growth engine is significantly shifting towards the AI application business, the essence of which is the production, delivery, and monetization of Scenario Tokens. Examining a longer timeframe reveals a more complete trend. According to the company's 2025 annual report, full-year revenue was 2.818 billion yuan, up 80.8% year-over-year. The AI application business segment, accounting for 52.8% of revenue, reached 1.487 billion yuan for the full year, a 76.5% increase. Revenue from precision marketing services was 1.331 billion yuan, up 85.8%. Adjusted net profit was 152 million yuan, growing 91.3%. This signifies that MARKETINGFORCE has moved beyond merely "adding AI on top of existing software" and is genuinely establishing Scenario Tokens as a major carrier of revenue and profit. Behind each AI revenue stream lies not just model calls or GenAI capability itself, but enterprise clients paying continuously for business outcomes, operational results, and Scenario Tokens within specific marketing, sales, and customer service scenarios. For capital markets, this transition from "selling tools" to "selling results" is more significant than单纯的 high growth figures alone.
The product foundation shows how MARKETINGFORCE encapsulates Scenario Tokens through platform capabilities. Crucially, the company concurrently launched GenAI OS, AI-Agentforce Intelligent Middle Platform 3.0, KnowForce AI Knowledge Middle Platform, and Data-Agent Business Analysis Master in Q1. This product matrix covers enterprise-grade AI Agent application clusters and AI Agent development/management platforms, integrating key links such as model fusion, computing power scheduling, knowledge governance, and agent collaboration, spanning core scenarios like marketing, sales, customer service, and R&D. In other words, MARKETINGFORCE aims to sell not point-based AI tools, not isolated GenAI functions, nor isolated Model Token calls, but a complete set of Scenario Token production capacity that can be embedded into business processes and ultimately billed based on results. Within this, the AI-Agentforce Intelligent Middle Platform handles the automated production of Scenario Tokens, the KnowForce AI Knowledge Middle Platform is responsible for infusing Scenario Tokens with industry knowledge, and GenAI OS serves as the operational foundation for Scenario Tokens. This "GenAI OS + AI-Agentforce + KnowForce AI + Data-Agent" product matrix is fundamentally not about selling tools, but about building a complete capability to encapsulate the value chain from Computing Tokens and Model Tokens to Scenario Tokens. This is also the core differentiation between MARKETINGFORCE and general model providers or point-solution AI tool vendors.
GEO represents the most representative implementation entry point for MARKETINGFORCE's Scenario Tokens. This is why GEO deserves specific attention. GEO (Generative Engine Optimization) is one of the more commercially mature, clearly demanded, and highly capital market-watched segments within MARKETINGFORCE's current Scenario Token business. In the search era, companies competed for keyword rankings; in the AI search era, they compete for the "right to be recommended" within answers. As the power of brand exposure shifts from SEO to the recommendation logic of large models, the truly valuable element is no longer a single piece of content, but the entire set of Scenario Token capabilities formed around "being seen, understood, trusted, and recommended by AI." MARKETINGFORCE positions GEO not as a monitoring tool, nor merely a content generator, but as a full-link operating system covering "insight - production - distribution - tracking - attribution." It corresponds not to minor adjustments within old marketing budgets, but更像 to a new type of infrastructure for the AI search era; and the ultimate billing, commercial, and charging logic for this new infrastructure is precisely the practical reflection of Token economics in the AI application layer: the Scenario Token. From this perspective, GEO is not a peripheral product but a key gateway for MARKETINGFORCE to translate its GenAI capabilities, AI-Agentforce capabilities, and KnowForce AI capabilities into tangible business outcomes, brand recommendations, and business growth. Whoever helps enterprises consistently secure the "right to be recommended" in the AI search era has a greater opportunity to convert Model Token consumption into high-value-added Scenario Token revenue.
Regarding the moat, MARKETINGFORCE possesses at least two layers of advantage that jointly support the barrier around its Scenario Token capability. The first layer is data and models. The company has served over 210,000 enterprise clients, covering 721 sub-sectors. This real marketing data, in turn, trains the Tforce vertical large model, making content generation more aligned with commercial scenarios and AI citation preferences. General model vendors can output Model Tokens but cannot naturally output high-quality Scenario Tokens because they lack the industry data feedback loop, real business feedback, and knowledge accumulated within an enterprise operational context. MARKETINGFORCE's 210,000 clients are the fuel for the continuous iteration, optimization, and enhancement of its Scenario Token commercialization capabilities. The second layer is commercialization efficiency. For new businesses like GEO, the greatest challenge is often not product development, but customer education and conversion. However, MARKETINGFORCE's existing customer base itself provides the most direct conversion soil. In 2025, the number of MARKETINGFORCE's Key Accounts reached 1,609, a 105.5% year-over-year increase, with KA ACV rising 60.6%. These high-value clients are not only a source of revenue but also the best validation ground for Scenario Token demand and the most direct adopters of Token economics in the AI application layer.
Looking ahead, Scenario Tokens are set to become the new pricing anchor for the AI application layer. Therefore, the true basis for re-evaluating MARKETINGFORCE lies not solely in the 110.5% Q1 growth rate, but in its position at a critical juncture of AI value chain migration: as Computing Tokens and Model Tokens gradually become low-margin commodities, the high-value ground is shifting towards Scenario Tokens. Whoever can translate underlying capabilities into business outcomes that clients are willing to pay for long-term has a greater chance of commanding influence in the AI application layer and mastering the pricing power of Token economics at the application level. In terms of billing models, MARKETINGFORCE is transitioning from traditional subscriptions towards pay-for-performance and usage-based billing. Products like GEO and EVA have already pioneered a hybrid "consumption + effect" charging model. This model essentially uses the Scenario Token as the unit of account—clients pay for each "recommendation," each "qualified lead," and each quantifiable business outcome, rather than merely for software licenses or seats. Leveraging its industry data accumulated from 210,000 clients, the Tforce vertical model, and the AI-Agentforce Intelligent Middle Platform, MARKETINGFORCE has already established a systematic first-mover advantage in the Scenario Token competition. For capital markets, the true focus for MARKETINGFORCE is not just the high growth of its AI application business per se, but its translation of previously abstract industry concepts like GenAI, Token economics, and Scenario Tokens into verifiable revenue, profit, and business results. It is therefore reasonable to infer that MARKETINGFORCE is positioning itself to capture the "Scenario Token" high ground first. For investors seeking to understand the investment logic of the AI application layer, tracking the growth rate of MARKETINGFORCE's GEO business, KA client conversion rate, and the penetration of the "consumption + effect" billing model may be more meaningful than单纯 tracking Computing Token prices or Model Token parameters. Because within the value chain of Token economics, the Token closest to the business outcome—the Scenario Token—is the most scarce and valuable.
Comments