Why 90,000 People Queue 4 Hours for This Chinese AI Video Tool

90,985th in Line—And Still Waiting

When a tech journalist tried to generate a video with ByteDance's Seedance 2.0 this week, the app delivered brutal news: You're number 90,985 in queue. Wait time? Four hours for a five-second clip.

Two hours later, the estimate had grown to six hours. Yet people keep joining the queue.

Why? Because China's creative industry is calling it revolutionary. Feng Ji, founder of the studio behind global hit Black Myth: Wukong, said he was "deeply shocked" by the model's capabilities. Pan Tianhong, a video creator with 15 million followers, declared it "thinks like a director."

But here's the catch: Most of the world can't even try it. ByteDance restricts access to users of its Chinese AI apps—Doubao, Jimeng, Xiaoyunque—creating such demand that people are reselling ByteDance accounts to eager overseas adopters.

The Great AI Divide: Hollywood vs. Beijing

The cultural split is striking. While Hollywood treats AI video generation with suspicion, China's entertainment industry is embracing it openly.

Jia Zhangke, a Cannes-winning director, recently "collaborated" with Seedance 2.0 to create a five-minute film featuring AI avatars of himself remaking scenes from his movies. "I don't worry about whether technology will replace movies," he wrote. "What matters is how people use the technology."

Meanwhile, Disney, Netflix, and Paramount have all sent ByteDance cease-and-desist letters alleging copyright infringement. The contrast couldn't be starker: Chinese creators are experimenting; Hollywood lawyers are mobilizing.

Afra Wang, who closely tracks US-China AI developments, notes the irony: "China hasn't produced any decent AI coding tool—Chinese developers all depend on Claude or Codex. But when it comes to video AI, China is miles ahead of the US."

Why China Leads in Video AI

The answer lies partly in regulatory philosophy. China's looser intellectual property protections have created an environment where AI models can train on vast amounts of content without immediate legal challenges.

When Pan Tianhong discovered Seedance 2.0 could mimic his exact speaking voice—without his explicit consent or compensation—he shrugged it off as "probably buried in some terms and conditions I agreed to."

This permissive approach has downsides. Social media is already flooded with AI-generated content featuring copyrighted characters: Wolverine fighting Hulk, Tom Cruise battling Brad Pitt, even a dance-off between Michael Jackson and Hitler.

The Infrastructure Reality Check

Despite the hype, Seedance 2.0 faces serious constraints. ByteDance estimates a 15-second video costs over $2 to generate, and the company lacks sufficient GPU resources for widespread adoption.

Users have developed workarounds: requesting videos after midnight when queues are shorter, generating shorter clips, even renting premium accounts from others. Some wait hours only to have their 99% complete video rejected by content filters, sending them back to the queue's end.

The Legal Reckoning

As Seedance-generated content spreads globally, the legal risks multiply. The model's ability to create convincing footage of copyrighted characters and real people raises questions that China's domestic market hasn't fully confronted.

ByteDance's silence on the cease-and-desist letters suggests the company is still figuring out how to scale globally while managing intellectual property concerns that barely register in its home market.