The Rise of AI-Driven Video Generation Tools in China
A New Era in Filmmaking
Artificial intelligence (AI)-driven video generation tools have swiftly transformed into a significant investment focus for major Chinese tech firms aiming to diversify their revenue streams. Following the groundbreaking release of OpenAI’s Sora, a text-to-video model in February 2024, several Chinese companies have accelerated the development of their competitive offerings. This technological evolution allows for the creation of short films and video series through AI, redefining content creation capabilities.
Kuaishou’s Strategic Shift
Kuaishou Technology, diverging from the industry’s chatbot development trends, has concentrated its efforts on large language models (LLMs) tailored primarily for its short video app and advertising services. Senior Vice President Gai Kun has expressed concerns about generative pre-training transformers, pointing out that they face obstacles in achieving robust reasoning capabilities. This challenge complicates the quest for artificial general intelligence (AGI) and has pushed Kuaishou to explore AI video generation as a promising growth avenue.
Investment in Video Generation Models
To capitalize on this opportunity, Kuaishou has committed substantial resources to foundational video generation models. Over two years, the company has built an extensive computing cluster capable of training models comprising trillions of parameters. As announced in August 2024, Kuaishou introduced Kling AI, which, as of its 2.1 release, can generate five-second 1080p videos. Impressively, Kling AI generated revenue of 150 million yuan (approximately $20.8 million) in the first quarter of 2025, marking a notable addition to the company’s revenue portfolio.
ByteDance’s Competitive Edge
Similarly, ByteDance has shifted its focus toward video generation. The company has restructured its leadership to allow former Douyin CEO Zhang Nan to steer AI projects within its video-editing division. Under Zhang’s leadership, computational resources were significantly augmented to develop Jimeng AI, a tool inspired by OpenAI’s Dall-E 2, designed to create images based on textual prompts. In September 2024, ByteDance expanded Jimeng AI with five additional video generation models that are now managed by the company’s Seed department for both internal and external applications.
Enhancing User Experience
In a bid to enhance user experience, Jimeng AI integrated DeepSeek’s R1 reasoning model in March 2025, improving the detail in users’ text prompts. As a result, Jimeng AI reported an impressive 8.93 million monthly active users, substantially outpacing Kling AI’s 1.8 million, according to data from QuestMobile.
Comparative Strengths of AI Tools
Video creators have observed that each AI tool possesses unique strengths: Kling AI is lauded for its ability to accurately depict character nuances, while Jimeng AI stands out with more fluid scene continuity and character movement. This burgeoning sector has paved the way for a creator economy, where tiered subscription systems were quickly instituted following the launch of Kling AI and Jimeng AI, providing a range of credit packages.
Supporting Creators and New Opportunities
Kuaishou has initiated programs to assist creators in securing contracts, while ByteDance has actively promoted campaigns to distribute Jimeng credits, fostering the development of micro-drama content. These initiatives are translating into new opportunities, especially in the realm of e-commerce-linked content creation, enhancing revenue prospects for both companies.
Challenges Ahead for AI Video Generators
Despite the impressive advancements in AI-driven video generation, these tools cannot yet produce seamless, feature-length films autonomously. Filmmakers such as Wei Lian acknowledge a pressing need to utilize multiple tools—like Jimeng AI, Kling AI, PixVerse, and Pika AI—to leverage each platform’s unique advantages. Notably, a dominant solution has yet to emerge in the market, indicating room for growth and innovation.
Future Prospects: Multimodal Models
The future of AI-generated video content appears promising, with expectations for further improvements as multimodal models continue to evolve. These advanced models, which integrate text, images, audio, and video, are anticipated to enhance the realism and complexity of generated videos, offering filmmakers and content creators new possibilities.
Emerging Competitions in the Creator Economy
With the launch of these advanced video tools, the dynamics of the creator economy are changing at a rapid pace. As Kuaishou and ByteDance continue to compete for supremacy, other players in the tech sector may also enter the fray, intensifying competition and accelerating developments in AI-driven video technology.
An Increasingly Accessible Market
The subscription models employed by Kuaishou and ByteDance make these AI video generation tools more accessible. By providing tiered offerings, they cater to diverse creators—from amateurs to professionals—facilitating broader participation in the burgeoning content creation ecosystem.
User-Focused Innovations
As these tools evolve, user feedback will be critical. Both Kuaishou and ByteDance are likely to invest further in R&D to fine-tune their offerings, prioritize user experience, and address the limitations currently present in AI-generated content.
Collaboration vs. Competition
The rapid expansion in AI video technology might also see collaborative efforts between companies, fostering partnerships that could enhance capabilities. Strategic alliances could enable technology sharing and pooling of resources, ultimately benefiting creators and consumers alike.
Conclusion: The Future is Bright for AI in Filmmaking
In summary, AI-driven video generation tools herald a new era in the filmmaking landscape. While challenges remain, investments by Kuaishou, ByteDance, and other tech firms indicate a robust commitment to developing technologies that can revolutionize content creation. As capabilities expand and evolve, the future looks bright for AI in filmmaking, promising an exciting new chapter for both creators and consumers alike.
