CITIC SEC: AI enhances the market adaptability of leading gaming companies, recommending targets with opportunities for AI UGC platform characteristics.

CITIC SEC released a research report stating that AI will not simply dilute the value of the gaming industry; instead, it will shift the industrys value focus from "the ability to produce content" to "the ability to continuously create high-quality products, attract traffic, and realize commercialization." The gaming industry is transitioning from an era of "capacity constraints" to an era of "scarcity reassessment," with leading companies benefiting the most. The firm predicts that the value release of AI in the gaming industry chain will occur in two stages: in the short term, AI will enter the R&D pipelines of leading companies, enhancing the maintenance of evergreen products, the development of high-spec new products, and global trial-and-error capabilities; in the medium to long term, as model capabilities improve, token costs decrease, and platform infrastructure strengthens, AI is expected to expand content supply through AI UGC and create new experiences through AI Native, further enlarging market scale. The firm emphasizes recommending leading companies with comprehensive industrialization capabilities and those with opportunities in AI UGC platforms, and suggests paying attention to companies with content platform bases. The main points from CITIC SEC are as follows: Technological Evolution: Efficiency is first realized on the development side, while applications on the player side begin to scale. On the developer side, AI is transitioning from point tool trials to systematic production: the application of AI by large enterprises is moving from point tools to systematic production pipelines, while small and medium teams increasingly leverage general models to complete research, coding, and prototyping, with the barrier for individual creators gradually dropping to natural language. According to a survey by the Game Industry Committee, 86.4% of companies have applied AI in the production stage, but 73.7% are still in the "AI-assisted exploration" stage; currently, it remains that "humans decide what to do, and AI assists in completing tasks faster." On the player side, AI NPCs, AI teammates, and built-in UGC have already entered leading products; in the first half of 2026, about 15% of the top 100 mobile games incorporated AI gameplay, contributing 42.6% of the sample revenue; the AI NPC-related gameplay in "Game for Peace" has accumulated user experiences of 110 million, with a peak daily active user count of 17.7 million, indicating that AI has begun to transition from R&D assistance to real player experience. Short-term Opportunities: The non-mobile content market is expanding, and AI is enhancing the market acceptance capability of leading companies. By 2025, global PC gaming revenue is expected to grow by 12.0% year-on-year to approximately 305.2 billion yuan, with China contributing 42% of the global increase; during the same period, China's broad PC consumption is estimated to be around 113.4 billion yuan, an increase of approximately 13.8% year-on-year, with domestic client game revenue reaching 78.16 billion yuan, up 14.97% year-on-year. The firm expects that by 2028, domestic client revenue will reach approximately 109.4 billion, 117.2 billion, and 125.1 billion yuan under conservative, neutral, and optimistic scenarios respectively, with the neutral scenario indicating a growth of nearly 50% compared to 2025. The firm believes that the growth is mainly driven by increased supply of domestically-produced high-spec content and a revival of existing PC demand. AI does not directly create demand, but rather breaks the constraints of "qualityspeedmanpower," enabling the same organization to simultaneously maintain evergreen products, advance more new products, and execute global projects; at this stage, its value is more reflected in improvements in human efficiency and project acceptance capability, with the potential to gradually unit project costs and profit margins. Medium to Long-term Opportunities: The mature UGC flywheel has already completed validation, while AI native UGC and AI Native still face dual barriers in models and platforms. Currently, established UGC is mainly represented by all-in-one platforms like Roblox and in-game UGC from titles like "Party Animals" and "Game for Peace." The core commercial loop is not built from scratch by generative AI; rather, AI serves to reduce creation, review, and matching costs within existing systems. In contrast, AI game creation platforms like TapTap and the project code-named Craft, as well as AI interaction content platforms like Aippy and Douyin Interactive Space, are still in the product validation and ecosystem building stages, with AI Native mainly focused on narrative adventure, investigative reasoning, and character dialogue scenarios that involve information interaction, with no large-scale commercial hits yet. The firm believes that the core constraints arise simultaneously from model capabilities and platform capabilities: on the model side, AI UGC faces a dilemma between quality and cost; iterative dialogue can enhance completion, but token and cloud costs will rise as well. If costs are controlled, it may lead to staying in the realm of homogenized, low-completion prototypes; AI Native requires longer context and ongoing reasoning to maintain consistency in plot, characters, and world status, which will also incur higher costs, while controlling costs and compressing context risks leading to forgetfulness and confusion. On the platform side, UGC not only needs creative tools but must also address various aspects such as cloud hosting, publication review, recommendation distribution, IAA/IAP, creator settlement, and content governance. Numerous issues remain unresolved: a large number of homogenized works receiving average exposure may harm community experience, cold starts could make it difficult for high-quality works to be discovered, and AI games present more complex interest matching and quality assessments than short videos and traditional games. Thus, AI UGC and AI Native still need to wait for token cost reductions, model capability improvements, and platform infrastructure maturity. According to Gamma Data, it is estimated that once AI matures, it could bring incremental value of 53.3 billion to 84.6 billion yuan to the domestic gaming market. Landscape and Barriers: Short-term success relies on industrialization, while medium to long-term success relies on platformization. The core of short-term competition is not who first integrates AI tools, but rather who can stably embed AI in the R&D pipeline to consistently produce high-spec content that meets commercial release standards. Tencent and NTES have formed leading advantages in core categories, long-term operations, R&D talent, gaming data, and global product capabilities. In the medium to long term, as token costs decrease and model capabilities increase, the current cost-quality constraints are expected to gradually weaken, shifting industry competition focus to the breadth of layout and depth of platforms: the breadth of layout depends on whether companies can timely seize opportunities related to AI game creation, interactive content, AI NPCs and teammates, AI Native, and world models, while the depth of platform depends on whether they have a complete infrastructure for creation, operations, distribution, payment, sharing, and governance, along with relatively mature platform operation experience that can continuously translate new supply into user consumption and creator income. The firm believes that AI will not simply erase industry barriers but may instead concentrate value further among leading companies that possess technological layouts, high-quality content capabilities, and complete platform flywheels. Risk Factors: Growth of the PC and console market does not meet expectations; the performance of high-spec new products falls short of expectations; the efficiency gains from AI tools, commercialization of AI UGC platforms, and AI Native products do not meet expectations; reductions in inference costs are not as anticipated; cloud operation and content governance costs exceed expectations; risks related to copyrights, data security, minor protection, and content compliance.