China Merchants Securities has released a research report indicating that AI is progressively shifting from model capabilities and capital investment toward application and commercial deployment. Initial validation of revenue growth and efficiency gains has already emerged in areas such as marketing, short dramas, and imaging tools. The gaming sector has delivered standout performance, driven by overseas expansion, a recovery in PC games, and new product cycles, with leading companies also demonstrating robust cash flow and dividend capacity. The film and television industry faced overall pressure in the first half, but improved supply of quality content during the summer season has supported a marginal recovery in box office figures. Meanwhile, the publishing sector remains in a phase of weak demand recovery, where low-valuation, high-dividend stocks continue to present notable allocation value.
Where the focus lies
Investment in AI continues to ramp up, and monetization at the application level is beginning to accelerate. Internet leaders in Hong Kong have maintained steady core operations, with Tencent's second-quarter capital expenditure reaching RMB 52.8 billion, up 64% quarter-on-quarter. AI investment is progressively reflected in model capabilities, AI applications, cloud services, and efficiency gains in advertising and gaming. At the same time, AI is transitioning from "capability building" into actual business operations. In marketing, it now covers core processes such as content generation, ad placement, and operational optimization. For instance, Yidian Tianxia's AI covers roughly 50% of core execution tasks, boosting per-operator efficiency by over 40%. Short dramas and AI-driven comic formats are also beginning to generate clearer revenue contributions.
Gaming momentum persists with both growth and yield appeal
In the first half of 2026, domestic gaming market revenue reached RMB 188.45 billion, up 12.17% year-on-year, while overseas revenue from self-developed games grew 30.22%, reinforcing overseas markets as a key growth engine. Among key A-share gaming companies, profit growth exceeded 50% for Century Huatong, Giant Network, Kain Ying Network, and G-bits. Additionally, Giant Network and G-bits have dividend payout ratios nearing or exceeding 70%. Current valuations for leading players generally range between 10 to 13 times earnings, with fundamental growth, improved cash flow, and low valuations collectively strengthening the sector's allocation appeal.
Near-term pressure in film and publishing with AI-enabled growth ahead
Domestic box office faced overall pressure in the first half, though quality summer releases drove a market recovery. As of August 31, 2026, national box office revenue stood at RMB 28.2 billion with 717 million admissions, down 28% and 21% year-on-year, respectively. The AI short drama market is expanding rapidly. According to DataEye Research, the overall micro-drama market in China is projected to surpass RMB 121 billion in 2026, with the AI short drama and comic segment expected to exceed RMB 40 billion. In the first half of 2026, the Shenwan publishing sector recorded total operating revenue of RMB 57.174 billion and net profit attributable to shareholders of RMB 6.861 billion, declining 10.82% and 18.43% year-on-year on an aggregate basis. The larger decline in profit relative to revenue indicates continued operational pressure across the industry.
Advertising and content sentiment edges up with AI opening new scenarios
In the first half of 2026, advertising spending across all media rose 8.5% year-on-year, while the internet advertising market grew 7.1%, suggesting a rebound in advertiser budget intentions. More importantly, AI is advancing from content generation tools into consumer insight, campaign optimization, and agent execution, giving rise to new scenarios such as GEO and AI search marketing. In film and television, while the overall market remained under pressure, summer box office reached RMB 12.498 billion, up 4.45% year-on-year. Improved content supply is driving marginal sector improvements, with leading companies poised to recover first.
Large model competition shifts from parameter expansion to capability-efficiency balance, benefiting application layers
The prevalence of million-token context windows in domestic models is growing, with Coding, Agent, and native multimodal capabilities becoming primary iteration directions. API prices continue to decline with a tiered pricing structure emerging, lowering model usage costs and enabling downstream applications to expand call volumes and commercial scale. The publishing sector continues to exhibit weak demand recovery characteristics. Attention should focus on companies with stable textbook and teaching-aid fundamentals, gradually deploying AI education products, and those offering low valuations and high dividends.
Risk factors
Risks include AI technology development and commercialization falling short of expectations, industry regulatory and content compliance risks, intensifying industry competition, and fluctuations in macroeconomic conditions as well as advertising and entertainment consumption demand.