BTS choreographer Youngjun Choi's MVNT wants to be the AI Suno of K-pop dance

MVNT wants to turn K-pop choreography into a generative AI library

Youngjun Choi, the choreographer behind signature routines for BTS and Seventeen, is betting that artificial intelligence can give anyone with a smartphone the ability to create K-pop-style dances — and give working dancers a way to get paid for the moves they invent.

Choi has co-founded MVNT with Joon Jung, a former dancer turned tech and crypto professional who is now CEO. The startup wants to build a generative AI platform for dance in the same way Suno lets people generate songs. Users would upload music, and the system would produce a routine drawing on a database of dance choreography. The company has raised $200,000 in seed funding from Seoul-based Mashup Ventures and is now looking for $1 million in a pre-Series A round.

The idea grew out of a copyright problem. Choi, who has more than 800 K-pop routines to his name, says dancers are often cut out of the money their work generates because there is no system for tracking how choreography is used or licensing it. MVNT is designed to let dancers build portfolios and receive credit whenever their uploaded routines are used, with the database then powering the AI generation tools.

The platform is still early: about 1,100 users and 2,000 dance routines created so far. The first target audience is meme creators, with professional choreographers and businesses expected to come later. MVNT has already worked with Epic Games to build an AI music-to-dance tool using Epic's Unreal Engine.

MVNT's IP problem and the risk inside its Suno strategy

Why MVNT started with an IP problem, not an AI pitch

The founders' original motivation was economic, not technological. Jung says choreographers can have "tens of thousands of pieces of choreography out there, but there isn't a system through which they can benefit." The US copyright system makes that difficult: the US Copyright Office generally does not consider individual dance moves, short routines or social dances copyrightable. The 2022 dispute in which choreographer Kyle Hanagami sued Epic Games over Fortnite routines was settled, but it showed how thin legal protection can be. MVNT's portfolio-and-credit model is an attempt to create a private compensation system where law and industry practice have not provided one.

The Suno comparison: a tool that starts casual before going pro

MVNT is explicitly modeled on Suno's path: start with casual creators and meme makers, then move upmarket to professionals. If AI music platforms have shown anything, it is that a low-friction tool can grow a user base quickly, but converting clicks into sustainable royalties for underlying creators is harder. MVNT's early numbers — 1,100 users and 2,000 routines — are encouraging for a pre-product stage, but they are not yet proof that dancers will earn meaningful revenue from AI-generated derivatives.

What the Epic Games partnership signals

The collaboration with Epic Games is the most concrete sign of adoption so far. Epic's Ohchan Kwon called MVNT "one of the companies that's driving the entertainment industry forward." Embedding MVNT's technology in Unreal Engine gives the startup a distribution channel into animation and interactive content, and gives Epic a way to keep its tools relevant as generative media spreads. But a senior evangelist's endorsement is not a commercial contract; the two companies have developed tools together, and the business terms remain undisclosed.

Where this leaves the choreographer-vs-AI debate

Choi is careful to frame the technology as an assistant, not a replacement. Asked whether AI could ever make a routine like Seventeen member Hoshi's "Spider" — which Choi has called his magnum opus — he said the tool should open "new possibilities for choreographers' creative directions." That framing matters because the same argument can be used defensively: if AI tools are assistants, choreographers both contribute to the database and control the creative direction. Whether dancers trust that split will determine whether MVNT's database can attract the professional-grade content it needs.

What dancers and creators should know about MVNT's early platform

For working choreographers and dancers:

• MVNT is building its portfolio database now, with 1,100 users and 2,000 routines already uploaded. Dancers who want to establish credit on the platform before AI monetization scales should consider documenting and uploading routines early.
• The platform's stated plan is to pay dancers when uploaded choreography is used. Before committing, examine the revenue-sharing terms carefully; the company has not published payout rates or examples of dancer earnings.
• Legal protection remains limited: because the US Copyright Office generally does not copyright individual moves or short routines, a private credit-and-licensing system like MVNT may be the only practical route to compensation — which cuts both ways.

For creators and enterprises exploring AI-generated movement:

• MVNT is targeting meme creators first. Businesses with animation workflows should evaluate the Unreal Engine integration, which gives MVNT a route into interactive content even while the standalone app is still in development.
• The startup is seeking $1 million in pre-Series A funding after a $200,000 seed round. Treat current user metrics as early signals, not market validation.