Kling AI, Kuaishou's video generation tool, has become the world's largest mobile AI video creation platform with 7.8 million monthly active users, significantly outpacing OpenAI's Sora at 4.7 million. This shift represents a major turning point in the AI video generation market, driven by Sora's shutdown and Kling's mobile-first design that appeals to everyday creators rather than just professionals .

How Did Kling AI Capture So Much Market Share So Quickly?

Kling AI's explosive growth stems from a combination of timing, product design, and ecosystem advantage. When OpenAI announced Sora's shutdown, it created a massive gap in the creator community. Kling AI filled that void by offering what Sora couldn't: a seamless mobile experience that lets creators generate videos anytime, anywhere without navigating complicated web interfaces .

The platform's success rests on three core strengths that resonate with how people actually create content today:

• Mobile-First Design: Unlike web-based competitors, Kling AI prioritizes smartphone creation, matching the habits of short-form video creators who work on the go rather than at desks.
• Advanced Motion and Lighting: The algorithm has been trained extensively on real-world physics, enabling it to generate visually striking dynamic scenes with realistic lighting and natural movement patterns.
• Ecosystem Integration: Backed by Kuaishou's massive short-video platform, AI-generated content has a built-in distribution network where creators can immediately share their work.

The numbers tell the story. In March, Kling AI averaged 7.8 million monthly active users compared to Sora's 4.7 million, a significant margin . More impressively, the platform's weekly active users grew 4% week-over-week, reaching an average of 2.6 million users, suggesting sustained momentum rather than a one-time spike .

What Does Kling AI's Success Reveal About the Future of AI Video Tools?

Kling AI's dominance signals a broader shift in how Chinese AI companies approach commercialization. Rather than releasing cutting-edge models as open-source tools for developers to use freely, companies like Kuaishou, Alibaba, and Zhipu AI are now prioritizing proprietary, closed models delivered through cloud platforms and subscription services .

This represents a fundamental business model change. Alibaba recently launched three proprietary models, including the multimodal Qwen3.5-Omni, which can process text, audio, images, and video, but made them available exclusively through its cloud platform and chatbot website rather than releasing them openly . The company previously released the Qwen3-Omni as open-source, but the newer 3.5 version is closed, signaling a deliberate shift toward monetization.

"Building a flagship model with state-of-the-art performance is always the top priority. Releasing smaller models that are less competitive is not particularly important or urgent," stated Chujie Zheng, a Qwen researcher.

Chujie Zheng, Qwen Researcher at Alibaba

Zhipu AI's leadership echoed this sentiment. The company's Chief Executive noted that customers who initially planned to deploy open-source models on their own servers are gradually switching to using cloud-based APIs instead . This shift means creators and businesses no longer need to download and run models locally; they simply pay for access through the cloud, which is easier for users but more profitable for companies.

Kling AI exemplifies this strategy. The platform runs on non-public models, meaning Kuaishou maintains complete control over the technology while users benefit from a polished, consumer-friendly product . This approach contrasts sharply with the open-source movement that dominated Chinese AI development just a year ago, when companies like Alibaba and Zhipu AI released dozens of smaller models freely to build developer communities and global recognition.

Why Are Chinese AI Companies Abandoning Open-Source?

The shift from open-source to closed models reflects a maturation of the Chinese AI industry. When DeepSeek R1 launched in January, it sparked a wave of open-source releases that helped Chinese companies gain international credibility. On Hugging Face, a popular platform for sharing AI models, Qwen accumulated more derivative models in the developer community than Google and Meta combined, largely by releasing many smaller models with billions of parameters that developers could download and use freely .

However, that strategy served its purpose. Chinese AI firms have now achieved global recognition and no longer need open-source releases to prove their capabilities. Instead, they're focused on profitability. Profitable AI products in China, including Kuaishou's Kling AI, already operate on non-public models, demonstrating that the market rewards closed, proprietary approaches when the product experience is strong enough .

This transition has profound implications for creators. It means fewer free, downloadable AI models available to independent developers and smaller studios. Instead, the future of AI video generation will likely be dominated by cloud-based platforms where users pay subscription fees or per-use costs to access powerful tools. Kling AI's 7.8 million monthly active users suggest creators are willing to accept this trade-off if the product is easy to use and produces high-quality results.

The broader lesson is clear: the era of Chinese AI companies competing primarily on open-source contributions has ended. The next phase is about building consumer-facing products that generate revenue while maintaining technological leadership. Kling AI's meteoric rise shows that this strategy is working, at least in the video generation space.