Guosheng: VOBILE GROUP (03738) AI service commercialization accelerates; maintain "Buy" rating

Guosheng released a research report stating it is optimistic that VOBILE GROUP (03738) will continue to benefit from the expansion of AI content supply, and that the company's AI content business opens up new growth space. The bank expects the company to achieve revenue of HKD 3.71/4.91/6.50 billion in 2026-2028, YOY +29%/32%/33%; net profit attributable to shareholders of HKD 340/510/710 million, YOY +68%/52%/39%; corresponding PE of 22X/14X/10X, and maintains the company's "Buy" rating. Event: On September 3, the company announced that its subsidiary has entered into a strategic partnership and commercial agreement with Unified National Networks Sdn Bhd to establish a strategic cooperative relationship and obtain 25MW of data center capacity equipped for AI computing deployment, adopting liquid cooling and high power density technologies, with an additional 25MW expansion option attached. Guosheng's main views are as follows: Layout of Brunei AI factory secures 50MW of computing power, supporting DreamMaker business expansion On September 3, 2026, the company's subsidiary entered into a strategic partnership and commercial agreement with UNN, Brunei's national telecommunications network operator, to establish and operate an AI factory based on NVIDIA computing cores, securing 50MW of computing power (of which 25MW is reserved for subsequent expansion arrangements) to support the expansion of the DreamMaker platform and related services. According to the Chinese Embassy in Brunei, the UNN data center is planned to commence operations in 2028. The bank believes that this cooperation provides computing power deployment conditions for DreamMaker to expand its service scale, and is expected to enhance the company's ability to undertake AI content generation demand from film and television studios, independent producers, and content creators. On this basis, the company is expected to further provide copyright protection and monetization services to the same customers, strengthening the synergy between content production and its existing copyright business. 26H1 company revenue grew steadily, mainly due to continued growth in US business and expansion in mainland China business In 26H1, the company achieved revenue of HKD 1.81 billion, up approximately 24% year-on-year. Revenue growth was mainly due to continued growth in US business and expansion in mainland China business driven by strategic and commercial partnerships. By region: 1) US and other regions: revenue of HKD 930 million, up 27% year-on-year. The company's business integration with PEX progressed smoothly, the music business quickly formed revenue contribution and became an important driver for expanding new customers; 2) Mainland China: revenue of HKD 880 million, up 21% year-on-year. The company further deepened cooperation with leading short video, UGC, long video, music, and micro-drama platforms, and advanced the going-global of digital content. Both subscription and value-added services achieved growth By business: 1) Subscription services: revenue of HKD 720 million, up approximately 17.7% year-on-year. Growth was mainly due to new customers being primarily large platform customers, with API-type services constituting the main source of new revenue. Increased platform content supply, higher frequency of work releases, and growth in AI-generated and re-created content drove rising API call demand; 2) Value-added and other services: revenue of HKD 1.09 billion, up 29% year-on-year. Growth was mainly due to increased AI content supply, development of element-level secondary creation, deepened cooperation with leading customers, and DreamMaker generating AI-related revenue for the first time. AI content supply expands, demand for copyright identification and rights confirmation is expected to accelerate Video generation models achieved capability improvements in the first half of 2026. Content production that previously required large teams and long cycles can now be directly generated by small teams or even individual creators through models, and content supply has increased substantially accordingly. With the rapid growth of AI-generated and secondary-created content, copyright identification demand has extended from traditional copy-level comparison to element-level identification of character images, scenes, logos, voices, and music clips. AI-native content (such as AI micro-dramas) began to be included on a large scale in the group's monetizable asset management in 26H1, and the rights confirmation and commercialization process of related content continued to accelerate. The bank believes that the large-scale production of new forms of content such as AI micro-dramas is expected to drive growth in demand for services including content identification, copyright management, licensing and revenue sharing, and revenue settlement, and to accelerate the commercialization of related businesses. DreamMaker and Vobile MAX enter early commercialization, AI services begin to contribute revenue AI-related revenue driven by DreamMaker and Vobile MAX was generated for the first time in 26H1. DreamMaker platform computing power service revenue grew from zero to the tens of millions of US dollars scale. Vobile MAX aggregates the company's core service capabilities in rights confirmation, protection, and monetization. DreamMaker serves as the creator entry point in the AI era, and the content it produces continuously enters Vobile MAX for rights confirmation and monetization. The company also completed the issuance of its first film and television copyright cash flow RWA project under Hong Kong's compliance framework, further exploring the trading and monetization of digital content assets. In terms of AI tools, Data Mining Copilot increased each engineer's productivity by approximately 75%, and reduced the cost of establishing and maintaining related monitoring algorithms by approximately 50%; the initial application of Intelligent Target Engine in music content increased the average revenue per claim to approximately 3 times that of the original method. The bank believes that as related platforms and tools are further implemented, the company is expected to improve content identification and monitoring efficiency, reduce algorithm establishment and maintenance costs, and drive copyright claim revenue growth by optimizing monitoring target screening. Risk warnings: AI-related business commercialization falling short of expectations, changes in copyright regulatory policies, intensifying industry competition, large customer renewal or business expansion falling short of expectations, etc.