The Rise of Video Generation AI: Transforming China’s Startup Landscape
Introduction to AI’s Financial Boom
Since the beginning of 2023, the buzz surrounding artificial intelligence has predominantly spotlighted two distinct realms: agentic AI, best exemplified by Manus, and AI hardware, often represented by representatives like Plaud.ai. However, a quieter yet no less impactful revolution is taking shape in China, where video generation models are igniting a new wave of innovative AI startups.
Emerging Giants in Video Generation
In June of this year, Kuaishou’s Kling AI achieved a remarkable milestone, reaching an annual recurring revenue (ARR) of $100 million across both its app and web platforms. Meanwhile, emerging players like MiniMax’s Hailuo AI and Shengshu Technology’s Vidu have each crossed the $10 million ARR threshold exclusively through their web-based offerings. Several industry insiders have suggested that the actual revenue from subscriptions might even exceed these figures.
<script async src="https://pagead2.googlesyndication.com/pagead/js/adsbygoogle.js?client=ca-pub-9958505722835444"
crossorigin="anonymous">
<ins class="adsbygoogle"
style="display:block; text-align:center;"
data-ad-layout="in-article"
data-ad-format="fluid"
data-ad-client="ca-pub-9958505722835444"
data-ad-slot="6218723755">
Profitability in Video Generation
What is striking is that video generation is yielding profits at a stage where large language models (LLMs) have yet to stabilize their profitability. AIsphere, for example, announced that its video model, PixVerse, has generated enough subscription revenue to cover a significant portion of its operational costs, bringing the company closer to breakeven.
At the 2023 Beijing Academy of Artificial Intelligence (BAAI) Conference, Huang Weilin—ByteDance’s head of image and video generation—projected that top-tier video generation products could potentially hit the $100 million ARR mark this year, with the opportunity to escalate that figure to between $500 million and $1 billion by 2026.
<script async src="https://pagead2.googlesyndication.com/pagead/js/adsbygoogle.js?client=ca-pub-9958505722835444"
crossorigin="anonymous">
<ins class="adsbygoogle"
style="display:block"
data-ad-format="autorelaxed"
data-ad-client="ca-pub-9958505722835444"
data-ad-slot="6793438825">
Changing Perceptions and Initial Doubts
Just a year ago, models akin to Sora were deemed too expensive to operate and fraught with commercial uncertainties. Tencent reported that Wang Changhu, an industry leader previously with ByteDance and the founder of AIsphere, received advice against entering this field. MiniMax’s Hailuo faced skepticism when it launched in September 2024, challenged by the competition from Kuaishou and ByteDance.
Many investors viewed video models as unlikely to yield timely returns, fearing that various startups would be at risk of being overshadowed by established giants, similar to trends observed in the LLM sector.
Funding Struggles and the Landscape
Indeed, in 2024, numerous Chinese video AI startups grappled with funding difficulties, and many could not achieve a sustainable product-market fit. Notably, Luying Technology, supported by Redpoint China Ventures and Lanchi Ventures, was acquired late that year.
Yet, just within a year, AIsphere’s revenue trajectory shifted the tide of investor sentiment. One investor reflected, “I deeply regret my earlier stance. We collectively underestimated the profitability potential of video generation models.”
<script async src="https://pagead2.googlesyndication.com/pagead/js/adsbygoogle.js?client=ca-pub-9958505722835444"
crossorigin="anonymous">
<ins class="adsbygoogle"
style="display:block; text-align:center;"
data-ad-layout="in-article"
data-ad-format="fluid"
data-ad-client="ca-pub-9958505722835444"
data-ad-slot="6218723755">
Three Factors Driving Growth
The surge in video AI startups can be attributed to three interconnected factors:
1. Aesthetic Appeal Over Technical Perfection
While video generation models are still polishing their capabilities, users seem more lenient towards imperfections. This forgiveness may stem from the fact that this domain is more about aesthetic satisfaction rather than rigid accuracy. “Even modest variations in data strategies can lead to uniquely different video styles,” noted an investor. This diversity has fostered a richly varied market.
Many users have noted that Kling excels in producing food-related and mukbang-style content, likely due to Kuaishou’s extensive collection of short-form food videos.
2. Cost-Effective Global Expansion
Interestingly, Western users generally demonstrate a higher willingness to pay for AI-driven tools and adopt them more swiftly. This aspect has proven beneficial for Chinese video AI startups venturing into international waters. MiniMax faced hurdles with subscription pricing at home but discovered a more expansive audience overseas, ultimately achieving over $10 million ARR.
Moreover, financial constraints have spurred Chinese startups toward operational efficiency, allowing them to maintain a competitive edge in global markets. Both Hailuo and Vidu reportedly operate at one-tenth to one-sixth the cost of generating videos of comparable duration and resolution as those produced by Sora.
<script async src="https://pagead2.googlesyndication.com/pagead/js/adsbygoogle.js?client=ca-pub-9958505722835444"
crossorigin="anonymous">
<ins class="adsbygoogle"
style="display:block"
data-ad-format="autorelaxed"
data-ad-client="ca-pub-9958505722835444"
data-ad-slot="6793438825">
3. The Power of Virality
Platforms such as TikTok and YouTube have played a critical role in driving user growth. An employee from AIsphere disclosed that PixVerse experienced a marked increase in interest following the viral success of its “Venom” special effects, amassing over a billion views across multiple platforms. Special effects and filters have also significantly aided user acquisition for both Pika and Hailuo.
“As companies today strive to meet tech benchmarks, it’s no longer sufficient to simply keep pace,” articulated an investor. “A proactive approach to discover or even create scalable demand is essential.” In the evolving landscape of video creation, users seek more than basic productivity tools; they desire reach, visibility, and incentives.
Compelling Market Validation
Recent rankings underscore this trend. According to a January assessment by Andreessen Horowitz, Hailuo AI ranked 12th globally in user traffic, outpacing even leading entities like OpenAI’s Sora and Kuaishou’s Kling, ranked 23rd and 20th respectively. This indicates that the video generation arena remains dynamic, providing opportunities for startups to establish themselves—unlike the more consolidated realm of LLMs.
Wang noted that video model development is currently somewhere between the GPT-2 and GPT-3 stages, indicating that there are ample opportunities for new contenders to emerge and tackle existing challenges.
A Shrinking Window for Advantage
However, as more players develop similar self-reinforcing strategies, the first-mover advantage appears to be diminishing. Current leaders are under increasing pressure to sustain their momentum. In March 2024, Wang cautioned that those who fail to act swiftly might encounter significant obstacles. “If you haven’t secured enough funding, attracted users, or established a solid workforce and tech base during the initial phase, your future in this market becomes tenuous.”
Another AI-focused investor echoed this sentiment, stating that while the market hasn’t reached saturation, there’s a dwindling appetite among investors for newcomers—barring the rise of a groundbreaking player akin to DeepSeek.
Funding Disparities
It’s worth noting that video AI startups typically garner significantly less investment than their LLM-focused counterparts. Over time, this gap is expected to widen, with established names like Kling and Dreamina holding a technical edge because of their rapid iteration.
For those already entrenched in this sector, the message is clear: innovate quickly, or risk being eclipsed.
Conclusion: Navigating the Future of Video Generation AI
The unfolding landscape of video generation in China serves as a testament to the dynamic nature of the AI industry. As these innovative startups continue to gain traction, their focus on aesthetic appeal, cost-effective expansion, and virality underscores a significant shift that may dictate the trajectory of future technologies. For stakeholders and investors, understanding these trends is crucial to navigating the competitive terrain. The future of video generation AI is not just promising; it is evolving rapidly, and staying ahead is paramount.
<script async src="https://pagead2.googlesyndication.com/pagead/js/adsbygoogle.js?client=ca-pub-9958505722835444"
crossorigin="anonymous">
<ins class="adsbygoogle"
style="display:block; text-align:center;"
data-ad-layout="in-article"
data-ad-format="fluid"
data-ad-client="ca-pub-9958505722835444"
data-ad-slot="6218723755">
