ElevenLabs has released Music v2, a more powerful version of its AI music system that lets creators edit tracks section by section. The update, which launched on May 26, marks a shift from one-time music generation to something closer to a full production toolkit.
The new features allow users to switch genres mid-track, rebuild bridges without touching choruses, and regenerate specific sections through a process called inpainting. This level of control could change how brands, creators and developers approach music production. But it also creates new challenges for an industry already struggling to define how AI-generated music fits into existing licensing frameworks.
Precision editing replaces one-shot generation
Music v2 focuses on giving creators more control over their output. Instead of accepting whatever an AI prompt produces, users can now:
- Move tracks between different musical genres
- Edit specific sections while leaving others untouched
- Regenerate vocals, instrumentation and arrangements
- Create multilingual tracks with improved performance
For independent artists and small studios, this could cut the time between having an idea and producing a working demo. Brands might find it cheaper and faster than licensing stock music or hiring composers for every campaign.
ElevenLabs has also cut pricing by up to 50% for its API and up to 40% for self-serve creative customers. The price reduction suggests the company wants to drive adoption among users who need music for videos, games or apps but don’t consider themselves musicians.
Three platforms target different markets
Music v2 now powers three different ElevenLabs products, each aimed at specific user groups. ElevenMusic gives creators tools to make and remix tracks directly. ElevenCreative targets advertising and branded content teams. ElevenAPI lets developers build music generation into their own applications.
This split shows ElevenLabs isn’t just competing for musicians. The company is going after marketing teams, app builders and media companies that need licensed audio quickly. Success in these markets could be more valuable than individual creator subscriptions.
Licensing becomes more complex with editable music
ElevenLabs claims its model uses only licensed training data and produces commercially cleared outputs. The company says creators can use generated tracks without sync fees or clearance delays. These are strong commercial claims that will likely face testing from brands, agencies and rights holders.
The editing capabilities make licensing questions harder to answer. Traditional music rights cover recordings, compositions, sync licenses and mechanical royalties. They weren’t designed for systems that can rewrite verses, shift into different genres for a few bars, then return to new vocal textures.
When a user regenerates just the bridge of a song, is that a new output or a derivative work? How do you track attribution when a track gets modified repeatedly? These questions matter more as the technology moves from novelty clips to production workflows.
Industry pushes for clearer AI music rules
The timing is notable. Universal Music Group and TikTok recently renewed a licensing deal that includes commitments to remove unauthorized AI-generated music and improve artist attribution. While this doesn’t directly affect ElevenLabs, it shows where major industry players are heading.
ElevenLabs has tried to position itself differently from AI music companies facing lawsuits over training data. The company has built licensing relationships with rights groups and publishers, and frames ElevenMusic around a fully licensed model designed to keep artists and songwriters involved.
Music v2 makes this approach more important. The more flexible the editing tools become, the more pressure there is to define what was generated, what data influenced it, and who gets paid when the result is used commercially.
What this means for the industry
For creators, Music v2 could provide immediate value. It offers faster ways to test genres, produce demos, create background music and refine sections without starting over.
For ElevenLabs, the opportunity extends beyond music subscriptions. If Music v2 works reliably, the company can sell API usage and enterprise tools to customers that already spend heavily on content. This puts ElevenLabs closer to agency budgets, streaming platform needs and software company requirements.
The main risk is that licensing clarity falls behind product capabilities. If rights holders accept the consent-first model and agree on workable accounting systems, Music v2 could become serious commercial infrastructure. If they don’t, the flexibility that makes the tool appealing could become a legal pressure point.
The next phase of AI music won’t be decided only by audio quality. Success will depend on whether the industry can make business rules as adaptable as the music generation technology itself.
