Massive capital inflows fuel rapid iteration and commercialisation of China’s AI video model sector

China’s AI video large model track is witnessing vigorous capital activity, with leading domestic firms securing successive large-scale financing rounds alongside accelerated technological upgrading. Four specialised AI video enterprises have registered nearly RMB 30 billion in new financing over the past three months, marking robust market confidence in the sector’s industrial prospects.

Keling AI has set a new global financing record for AI video model firms with a recent financing round of nearly USD 3 billion, pushing its post-investment valuation close to USD 18 billion. The round is jointly led by CPE Yuanfeng, Guofang Venture Capital, BlueFive, Tencent, Zhongguancun Science City Fund and CITIC Securities, with participation from dozens of top-tier investment institutions. Leading industrial investors including Alibaba Cloud and Baidu, alongside major cultural and entertainment industrial players such as Huace Film and TV and Mangguo Industrial Investment, have also joined the financing lineup.

Multiple key financing deals have been completed across the sector since June. In July, general world model firm Shengshu Technology closed a USD 500 million financing round led by Alibaba Cloud, with national funds, industrial capital and prominent venture capital institutions participating. Also in July, AI video generation firm Aishi Technology finalised its overall Series C financing totalling RMB 2.98 billion, with the Series C+ round backed by Alibaba. In June, Yanyu Technology, operator of the AI video creation platform LibTV, secured nearly USD 300 million in Series B+ financing, achieving a post-investment valuation exceeding USD 2 billion.

Major internet platforms are actively engaging in the sector as both investors and technology developers. Tencent, Alibaba Cloud and Baidu have all participated in Keling AI’s latest financing. Alibaba has additionally led investment in Shengshu Technology’s Series B round and Aishi Technology’s Series C+ round. Domestic tech giants are speeding up independent model iteration, with ByteDance launching the upgraded Seedance2.5 audio and video generation model, Alibaba releasing the Wan2.7-Video model and Tencent Hunyuan unveiling the HunyuanVideo1.5 video generation model.

The sector’s financing boom signals a clear shift in market valuation logic, moving beyond conceptual speculation to substantive commercial evaluation. The industry has completed preliminary technological verification and entered a critical phase of industrial implementation. Industrial competition is increasingly centred on the efficiency of converting technological capabilities into sustainable commercial value, with scalable operational capacity set to define market leadership.

Technological progress and commercial monetisation are advancing in tandem to accelerate industrial validation. ByteDance’s Seedance2.5 model has achieved breakthroughs in four core capabilities compared with its earlier iteration, extending maximum single video generation duration to 30 seconds. AI-generated short drama production has emerged as a highly viable application scenario, with production costs drastically reduced from the traditional RMB 300,000–500,000 per project to RMB 30,000–50,000 while maintaining commercially viable quality standards. High-value commercial orders for AI virtual drama characters have further demonstrated the model industry’s market potential.

Operational data from leading firms verifies strong commercial growth momentum. Keling AI surpassed 100 million global users in early June 2026, representing a 67% increase from 60 million users recorded at the end of 2025, with service coverage extending across 224 countries and regions. The enterprise generated over RMB 650 million in operating revenue in the first quarter of 2026, marking a year-on-year surge of more than 300%. Its annual recurring revenue reached nearly USD 500 million in March 2026, a fivefold increase from USD 100 million in March 2025.

Kunlun Tech’s AI short drama business has maintained rapid expansion. The company’s short drama and AI short drama platforms recorded monthly gross revenue exceeding USD 48 million by late March 2026, with annual recurring revenue surpassing USD 570 million and monthly active user volume ranking first among overseas short drama platforms. Annual recurring revenue for the firm’s AI short drama business exceeded USD 700 million by the end of June 2026.

Core technical bottlenecks restricting AI video generation have been largely resolved, with consistent character rendering, long-sequence narrative logic and frame coherence fully optimised to support reliable paid commercial services. Industry players emphasise that the AI video sector requires long-term operational investment, with sustainable growth built on continuous application value creation and positive circulation between technological research and commercial returns.

Growing industrial prosperity is accompanied by clearer operational barriers and market differentiation. Elevated computing power costs, fluctuating consumer user retention rates and evolving copyright compliance requirements remain universal challenges for industrial participants. Sector competition has developed into comprehensive rivalry covering computing resources, data accumulation, technical talent reserve and strategic layout, with pure capital investment unable to sustain long-term competitive advantages.

According to a research report from CITIC Securities, the AI video generation tool sector will follow a developmental trajectory similar to the global game engine industry. Rigid technical and operational barriers will narrow the number of viable long-term market players, with the future industrial landscape likely to form an oligopolistic pattern featuring three to four differentiated leading enterprises.

Current industrial challenges are integral to the sector’s maturity process, offering enterprises opportunities to optimise operational mechanisms and build core competitiveness. World model technology is poised to become fundamental infrastructure for the global film, television and gaming industries over the next three to five years, reshaping traditional content production workflows. AI video tools will evolve from efficiency-enhancing auxiliary instruments into intelligent creative systems capable of independent script understanding and lens design, driving fundamental upgrades across the entire content industry chain.