Author｜Yang Licheng

Editor｜Chen Xiaoran

On June 5, Hong Kong-listed veteran shoe company C.banner Holdings (01028.HK) announced a 300 million yuan controlling stake in AI data service provider Benyuan Zhishu. This seemingly incongruous cross-border M&A unveils the revaluation of data value in the upstream AI industry chain.

On the evening of June 5, C.banner Holdings issued an announcement stating its intention to acquire approximately 90.91% of Benyuan Zhishu's equity through its wholly-owned subsidiary Shanghai Qiandu for a total consideration of 300 million yuan. The transaction will be completed in two steps: 150 million yuan for the acquisition of 83.33% of existing shares and an additional 150 million yuan for the subscription of new shares.

After the transaction, Benyuan Zhishu will be included in C.banner Holdings' consolidated financial statements, and the company will officially form a dual-main-business structure of "footwear manufacturing and sales + artificial intelligence data services."

As of the close of trading on June 8 (Beijing Time), C.banner Holdings' stock price stood at HK$1.44 per share, down 16.76%, with a total market capitalization of approximately HK$3.6 billion. Despite the day's price correction, the company's stock price has still surged by over 100% in the past three months, reflecting market expectations for this cross-border transformation.

The transaction is ingeniously designed. C.banner Holdings will pay the entire consideration using internal resources, maintaining the footwear business's net cash position without diluting existing shareholders' equity.

More critically, C.banner Holdings explicitly stipulates that after the transaction, Benyuan Zhishu will retain its independent brand, operational autonomy, and a complete data isolation mechanism. The founding team will still hold the remaining 9.09% equity, and the core management will remain unchanged.

C.banner Holdings also commits to being "just a data supplier, not a large model developer," refraining from intervening in Benyuan Zhishu's daily operations or competing with its downstream clients.

This "financial holding + business isolation" model precisely addresses the most sensitive neutrality issue in the AI data services industry.

In the AI industry, once a data service provider is controlled by a large model firm, its competitors often switch providers due to data security and trade secret concerns.

As a traditional consumer enterprise completely unrelated to AI model businesses, C.banner Holdings becomes the best safeguard for Benyuan Zhishu's neutrality, allowing rival tech giants to confidently continue entrusting their data business to the same supplier.

Founded in 2015, Benyuan Zhishu is an undisputed "invisible champion" in China's AI data services sector.

During the early stages when the industry primarily focused on computing power and model architectures, Benyuan Zhishu chose a less-traveled path: avoiding low-threshold, highly competitive general data annotation and instead specializing in highly technical, non-standardized professional data services.

Benyuan Zhishu's core business involves transforming human expertise and cognitive experience into high-quality data trainable and verifiable by large models.

It has established a resource pool of hundreds of thousands of experts covering over 20 fields, including mathematics, coding, law, and healthcare, providing expert-level annotation services.

It also constructs evaluation datasets to measure large models' true capabilities, offering iterative performance verification for leading firms.

Additionally, it provides virtual environment reconstruction data for world model development and collects real-world interaction data for embodied AI companies.

After eleven years of deep cultivation, Benyuan Zhishu has built formidable competitive barriers.

Currently, Benyuan Zhishu operates ten delivery centers nationwide, possesses self-developed annotation, quality inspection, and crowdsourcing management platforms, holds comprehensive data security and compliance certifications, and is one of the few domestic suppliers capable of providing full-scenario data services for large models, world models, and embodied AI.

Benyuan Zhishu's clients include over 80% of China's leading large model firms, top-tier internet platforms, and leading embodied AI companies, serving as an exclusive supplier for multiple high-value data categories.

Against the backdrop of the AI industry's widespread "burning money for scale" approach, Benyuan Zhishu's profitability is particularly rare.

The announcement reveals that Benyuan Zhishu achieved revenue of 146.5 million yuan in 2024 with a net profit of 7.1 million yuan; revenue grew to 156.2 million yuan in 2025 with a net profit of 11.1 million yuan, up 56.3% year-on-year; and revenue continued its strong growth in the first five months of 2026.

Since its inception, Benyuan Zhishu has never relied on substantial external financing, achieving sustained profitability and business expansion solely through operational cash flow.

Behind this profitability lies Benyuan Zhishu's precise strategic positioning. It has avoided the price war red ocean of low-threshold general data annotation and focused on high-value-added professional data services. Benyuan Zhishu's business layout (business layout) precisely aligns with the core nodes of industry demand upgrades.

C.banner Holdings' seemingly incongruous acquisition precisely aligns with the underlying logical shift occurring in the global AI industry.

As computing power resources become increasingly accessible and model architectures converge, high-quality, verifiable professional data has emerged as the core bottleneck restricting AI's advancement to higher stages.

Computing power can be procured, and algorithms can be replicated. However, the ability to scale the professional judgment of human experts and real-world interaction feedback into model-learnable data cannot be rapidly duplicated by capital, representing the most irreplaceable core link in the AI industry chain.

Global capital has already voted for this logic with real money. In June 2025, Meta invested $14.3 billion in U.S. data annotation leader Scale AI, valuing it at approximately $29 billion, redefining AI data services from a past cost item to a strategic asset in the AI era.

In the domestic market, according to iResearch data, the scale of China's AI foundational data services market surpassed 102 billion yuan in 2025, with a compound annual growth rate of approximately 21.8%.

2026 is widely regarded by the industry as the "year of data scaling" for embodied AI. As AI moves beyond screens into the physical world, demand for real-world interaction data will explode.

A frequently overlooked industry truth is that in actual embodied AI delivery, real data accounts for over 95% of data procurement; all simulated data must ultimately be calibrated and verified against real-world data.

Players engaged in source-level real data collection and delivery are destined to secure a "foundational" position in the future AI industry.

For C.banner Holdings, this acquisition represents a crucial move in seeking a second growth curve. The company's 2025 annual report shows full-year revenue of 1.213 billion yuan, down 12.8% year-on-year. While the footwear business provides stable cash flow, growth is sluggish amid fierce industry competition.

After entering the AI data sector, assuming industry average growth rates, if Benyuan Zhishu maintains a 30% compound annual growth rate, its revenue could exceed 1 billion yuan by 2030 with net profits surpassing 200 million yuan, contributing significant profit increments to C.banner Holdings and driving the company's valuation system to restructure from a traditional consumer firm to a "consumer + tech" dual-main-business enterprise.

Multiple potential risks remain that investors should note. The pace of AI industry commercialization is uncertain, and delays in large models' and embodied AI's market adoption could directly drag down data service demand.

Increasing industry entrants and cross-border layouts by leading tech firms may divert existing orders and compress profit margins.

Technological disruptions also pose challenges, as AI auto-annotation and synthetic data mature, traditional manual annotation businesses face substitution pressure.

The two companies have vastly different corporate DNA, with significant disparities in management styles and team cultures. Inadequate integration could easily lead to the loss of core talent.

In the long run, China's AI industry has shifted from a computing power race to a data race, with upstream data service providers possessing core technological barriers and stable profitability set to benefit continuously.

For traditional industrial firms seeking AI layout (layout) opportunities, the collaboration between C.banner Holdings and Benyuan Zhishu offers valuable insights. By avoiding homogeneous model competition and targeting high-barrier segments of the industry chain while leveraging a stable main business to empower tech operations, they can achieve synergistic progress between industry and technology.