500 million market value fixed increase, cooperation with MetaX: MARKETINGFORCE bets on the capacity base for AI scaled delivery

The competition at the AI application layer is shifting from "whether there are scenarios" to "whether it can be scaled for delivery"; the valuation logic is also transitioning from "whether there are models and computing power" to a more realistic question—what business flows has AI actually entered, and can it form sustainable billing and customer ROI.

On May 20, MARKETINGFORCE (02556.HK) officially completed a highly anticipated new share subscription, raising a net amount of approximately HKD 499.7 million. This funding did not come like traditional placements: the pricing was basically consistent with the closing price on that day, representing a standard market price issuance with no discount. This indicates that the introduced partners are not merely arbitrage funds chasing short-term price differences, but rather strategic capital that recognizes the company's long-term positioning. The announcement further clarified that 100% of the net proceeds will be directed towards the construction and operation of intelligent computing infrastructure.

Shortly after, on May 22, the company announced a strategic cooperation with domestic GPU manufacturer MetaX, incorporating the fully domestically produced Xiyun C600 chip into its computing power matrix. Within a week, with two consecutive moves, MARKETINGFORCE laid out a clear card in front of the market: it is no longer just about the AI concept, but about pushing the "full-stack Token factory" from concept into the underlying delivery capacity required for real business processes.

Supplementing capacity rather than speculating on concepts: The delivery logic behind MARKETINGFORCE's HKD 500 million financing

In the past, the market valued AI application companies based on cases, clients, and product matrices. However, when AI entities truly enter the high-concurrency, long-chain, and high-frequency business processes of enterprises, a more fundamental question arises: do you have enough stable, prioritized, and schedulable computing resources to support the transition from "single-point pilot" to "scaled delivery"?

This is precisely the underlying logic of this HKD 500 million market price private placement—this is a "capacity construction" aimed at AI delivery capabilities.

The funding directly targets capacity bottlenecks: GPU server procurement, networking, AIDC leasing, as well as software for AI large models and model management platforms. The goal is singular: to meet the continuously growing computing power demands of customer AI applications while enhancing the training efficiency of proprietary vertical models. More importantly, the announcement clearly emphasizes the need to "introduce strategic partners in the intelligent computing industry to further ensure the stability and priority of the group's computing power supply chain." This means that MARKETINGFORCE has upgraded intelligent computing resources from ordinary procurement costs to core production factors and has locked in supply chain priorities in advance.

The subsequent cooperation with MetaX further resembles building upon a "self-built foundation" to construct a diversified supply matrix of "general cloud computing power + domestically produced autonomous computing power." MetaX's Xiyun C600 has achieved full-process domestication from IP architecture to supply chain and has entered mass production, allowing MARKETINGFORCE to directly gain access to high-value markets in scenarios such as government affairs and energy, where there are rigid requirements for security and controllability.

At this point, MARKETINGFORCE's closed-loop logic of "computing power—middle platform—scenarios" is more complete: the HKD 500 million market price private placement builds an intelligent computing foundation, combined with diversified computing power from MetaX and others, answers the question of "where does computing power come from"; the AI-Agentforce intelligent entity middle platform and KnowForce AI knowledge middle platform complete the accumulation of industry cognition, and based on this, coordinate and schedule suitable intelligent entity models for process orchestration Answered "How to consolidate capabilities"; ultimately, in scenarios such as government enterprises, new energy, and retail, processing computing power into measurable and chargeable scenario Tokens, answering "Where does value go?"

Domestic Computing Power Entering Enterprises: AI Applications Entering a New Stage of "Result-Based Pricing" Delivery

If supplementing production capacity is "rooting down," then MARKETINGFORCE must also "result upwards." What the capital market will focus on next has long exceeded the actions themselves; the market and investors' attention will be more on whether it can produce more, more complete, and more valuable "scenario Token" samples.

At the foundational level, MARKETINGFORCE positions itself as a full-stack Token factory, building the GenOS AI native operating system, and relying on the scheduling and coordination of the AI-Agentforce enterprise-level intelligent agent platform and the industry knowledge Know-how of the KnowForce AI knowledge platform, ultimately landing on the practical execution level of Data-Agent business analysis masters and other scenario-based Token product matrices.

In this enterprise process that connects "computing power, models, enterprise knowledge, and business scenarios" to form a complete closed loop, the key role that truly creates value and makes enterprises continuously pay for it is the "result-type Token" that can directly enhance business efficiency after processing. This is vividly summarized by MARKETINGFORCE as "scenario Token," and its value is precisely the main line that MARKETINGFORCE attempts to prove. Thus, it can be seen that MARKETINGFORCE's product matrix does not stop at single-point tools but penetrates into the high-value business processes of enterprises:

The task unit of the solution assistant is a deliverable plan, compressing the manpower and time of repeated communication in the sales chain.

The task unit of the customer service auxiliary Agent is to complete an effective auxiliary conversation that includes script recommendations, knowledge retrieval, emotional judgment, and ticket distribution, with the core value being to enhance service consistency and repurchase conversion.

The task unit of the GTM Agent is a complete market entry or new product launch strategy, addressing the issue of strategy efficiency.

The task unit of the ToB customer service Agent is to handle a longer and more complete business process scenario covering technical parameters, quotations, contracts, pre-sales support, etc. Especially in scenarios such as manufacturing, new energy, and going overseas, these business scenarios within enterprises that are closer to real budgets often become high-value segments where "scenario Tokens" are first realized.

From a more macro perspective, these scenarios collectively illustrate a trend: AI is moving from individual tools into the complete processes within enterprises and is beginning to truly create value. Once enterprise intelligent agents enter a 7×24 hour continuous operation state, what corresponds behind it is continuous, stable, and predictable Token consumption and computing power demand. This also explains why MARKETINGFORCE must invest HKD 500 million to build its own intelligent computing foundation and why it needs to access MetaX's domestic GPUs.

In the first quarter of 2026, MARKETINGFORCE's AI application business revenue grew by approximately 110.5% year-on-year. After rapid growth, the problem facing MARKETINGFORCE is no longer "Is there AI business?" but whether it can deliver stably, at low cost, and with high quality From the complex processes of government and enterprise in Hefei's smart investment promotion, to the industrial operation scenarios of new energy customers, and then to the user operations in consumer retail and cultural tourism, these pieces are trying to build a ladder for MARKETINGFORCE to move towards an "enterprise-level AI application native platform."

Long-distance track: Three indicators that need to be validated next

Of course, this logic still needs continuous validation. What the market really needs to look at later is not the density of concepts, but the hard indicators used to assess the positioning of value layers:

Can the scale of scene Token consumption continue to grow: This represents whether customers are truly internalizing AI into their daily business flows;

Can AI application revenue and gross margin improve simultaneously: This reflects the scale effect of delivery capability and value creation;

Can customer ROI be replicated in more industries: This determines the ceiling of the "full-stack Token factory."

If these three points gradually hold true, the meaning of the "full-stack Token factory" will become clearer: it is not about making enterprises consume more Tokens, but about processing Tokens into business results that enterprises are willing to pay for continuously.

From the announcement on May 15 to the subscription confirmation on May 20, and then to the signing with MetaX on May 22, the series of moves made by MARKETINGFORCE may be laying a stable, controllable, and scalable delivery foundation track for the long-distance run of the "first stock of scene Tokens." The final result, however, still requires the market and investors to extend the observation time cycle to track and evaluate the ultimate realization results