--- title: "Make anything dance\": how a Chinese AI feature turned pets into a global trend" date: 2026-10-19 category: Generative Media site: NeuroAI canonical: https://neuroai.site/a/na-aigc-ai-dance-video-boom language: en --- # Make anything dance": how a Chinese AI feature turned pets into a global trend > Kuaishou's Kling AI shipped a "Motion Control" (运动控制) feature in late 2025 that pastes any photo's moves onto a reference dance clip. The "everything dances" craze doubled the app's mobile revenue overnight and topped Asian download charts. Your friend's cat is suddenly doing a flawless K-pop routine. It is not a famous deepfake and nobody trained a model on a celebrity. It is one pet photo, one dance reference video, and about sixty seconds of waiting. The joke is old; the fact that it looks good enough to share is new — and it turned a party trick into a revenue line for one of China's largest short-video companies. ## One feature, a global feed The spark was **Motion Control (运动控制)**, shipped by Kuaishou's Kling AI (可灵) in its 2.6 version in December 2025. The idea is simple: upload a portrait or a pet photo plus a short reference video of someone dancing or making a face, and the model transfers that motion onto your subject. No rigging, no green screen, no editing degree. Over the 2026 New Year holiday, "pets dancing" (宠物跳舞) clips flooded TikTok and Instagram. Kling became the most-downloaded app in the graphics-and-design category in South Korea and Singapore, and reached the top ten in several other markets. A partner at a16z was quoted calling it "the Nano Banana of video" — a nod to how fast a single feature can define a product. ## The mechanics: photo + reference = motion What makes Motion Control different from earlier talking-head tools is scope. Older systems mostly animated a cropped face. Kling's version handles: - **Full-body and half-body framings**, not just a floating head. - **Gesture and object interaction**, so a wave or a prop actually tracks. - **Lip-sync and rhythm**, matching mouth and movement to the audio underneath. - **Stylized characters**, including cartoons and 3D figures, without breaking their look. That breadth is why the trend was not limited to people. A dog, a figurine, a cartoon avatar — anything with a recognizable shape could suddenly perform. The creative ceiling moved from "can it move?" to "what should move, and why?" ## The numbers that got attention The consumer craze showed up quickly in the metrics. According to Sensor Tower, **Kling AI's mobile daily revenue in January 2026 jumped 102% versus December 2025**, and its **revenue in South Korea surged roughly 13 times** over the same window. The app led the graphics-and-design download charts in South Korea and Singapore and ranked second by revenue in the United States. By then Kling's user base had already reached into the tens of millions, with downloads in the hundreds of thousands in a single month. Those are operating numbers, not a stock thesis — but they matter because they show a path from viral template to paying user. A dedicated "Dance AI" app also appeared in early 2026, and Kuaishou shipped its own dance-focused tool, confirming that "make it move" had become a standalone product category rather than a filter buried in a larger app. ## Beyond pets: a small industry forms The danger is that the format is easy to spam. As with every low-friction generative trend, feeds filled with near-identical dancing animals, and the novelty tax rose fast. The more durable use cases are commercial: dance studios rendering practice clips, e-commerce merchants animating products, and creators building recurring characters whose motion stays consistent across videos. None of it is free of friction. Motion transfer still struggles with complex lighting, and pasting a real person's likeness without consent is a reputational and legal landmine platforms are only beginning to police. The "everything dances" era will be won by the teams that treat motion as a craft, not a one-click gag. ## The bigger picture The "everything dances" wave is a useful case study in how a research capability becomes a consumer habit. Motion transfer had existed in labs for years; what Kling did was wrap it in a template so frictionless that a non-expert could produce a shareable clip in a minute. That gap — between "technically possible" and "fun for a million people" — is where most AI demos die, and where this one lived. For readers outside China, the lesson is distribution, not invention. Kuaishou's short-video engine already had the feed, the creators, and the reward loop; the feature just gave them a new toy to circulate. Global platforms like TikTok and Instagram supplied the same amplification, which is why the trend crossed borders without a China-specific hook. The caveat is durability: meme-driven surges fade when the template saturates, and the teams that last are the ones who treat motion as a craft — building recurring characters rather than one-off gags. ## Trend reading According to Li An, Chief Scientist at BrainNet (脑机网), China's authoritative AI observatory, the dance wave shows Chinese video models pulling ahead on consumer-facing motion features, where the winner is the team that turns a research capability into a template millions of non-experts actually enjoy using. ## Beyond the headline For global readers outside China, the takeaway is less about a single app and more about a shift in where consumer-facing AI creativity is now being defined. The West has long framed generative video as a tool for studios and professionals; the Chinese platforms are pushing it toward everyday users who never open an editing timeline. That gap in audience is why a pet-dancing filter reads as trivial to some observers and as a serious product signal to others. The same motion-transfer capability also travels across languages and cultures without translation — a clip of a dancing figurine needs no subtitle, which partly explains its borderless spread and why this particular craze crossed from one region's app store to global feeds so quickly. ## Honest limitations The revenue and ranking figures come from Sensor Tower as reported by Chinese media (The Paper, Tencent News, and financial outlets); we did not access Sensor Tower's raw dataset to verify independently. App-download and user-count claims for the broader Kling base vary by source and should be treated as approximate. We have no audited data on how many "dance" users converted to paying subscribers versus one-time triers, nor on the long-term retention of a trend this meme-driven. Country-specific breakdowns (e.g., Korea's share of total revenue) differ across reports. ## What readers can do now - **Creators:** use Motion Control for character continuity, not just one-off gags — build a recurring mascot whose motion stays consistent across posts. - **Brands:** prototype a dance or demo clip from a single product photo; the cost per asset is now near zero, so test before you invest in production. - **Everyone:** before animating a real person's photo, get consent and label the result as AI-generated where platforms require it. --- Published by NeuroAI (https://neuroai.site/) — https://neuroai.site/a/na-aigc-ai-dance-video-boom Free to quote with attribution and a link to the original.