ElevenLabs has pushed its music model into more editable territory, and that makes Music v2 as much a licensing story as a product launch.
ElevenLabs released Music v2 on May 26, putting a sharper version of its AI music system behind tools for creators, developers and brands. The headline feature is not just better sound. It is control. A user can move a track from one genre to another, rebuild a bridge without touching the chorus, and shape a song section by section instead of accepting a single prompt output as the final product.
That matters because AI music is moving away from novelty clips and toward production workflows. ElevenLabs says Music v2 improves vocals, instrumentation, arrangements and multilingual performance, while also supporting inpainting that lets creators regenerate only a selected part of a track. In plain terms, the company is trying to make an AI song behave less like a finished file and more like an editable project.
As ElevenLabs explained in its launch materials, Music v2 now powers ElevenMusic, ElevenCreative and, soon, ElevenAPI. That split shows where the business is going. ElevenMusic gives creators a place to make and remix tracks, ElevenCreative points the technology at ads and branded content, and ElevenAPI opens the door for developers to embed music generation directly into other products. The model is therefore not just competing for musicians. It is chasing marketing teams, app builders and media companies that need licensed audio quickly.
The Product Leap Is Really About Control
Music v2 arrives with a practical promise: fewer throwaway generations and more precision. A creator can start with a mood, lyric or reference, then keep adjusting the structure until the track feels usable. For independent artists and small studios, that can shorten the distance between an idea and a polished demo. For brands, it could make custom audio cheaper and faster than clearing a stock track or commissioning a composer for every campaign.
The pricing move is part of the same strategy. ElevenLabs said it cut Music v1 and Music v2 pricing by up to 50% for ElevenAPI and up to 40% for ElevenCreative self-serve customers. That is a clear attempt to drive usage, especially among customers who may not think of themselves as musicians but still need music inside videos, games, training content or product experiences.
The sharper question is whether the market trusts the outputs. ElevenLabs says the model is trained only on licensed data and cleared for commercial use. It also says creators can use generated tracks without sync fees or clearance delays. That is a strong commercial claim, and it is exactly the sort of claim that will now be tested by brands, agencies and rights holders looking for clear accountability.
Licensing Gets More Complicated When Songs Become Mutable
ElevenLabs has tried to position itself differently from AI music companies that faced lawsuits over training data. Its earlier push into music leaned on licensing relationships with industry partners, including deals reported around ElevenMusic with rights groups and publishers. Music Business Worldwide recently noted that ElevenLabs has framed ElevenMusic around a fully licensed model and a commercial approach designed to keep artists and songwriters involved.
Music v2 makes that approach more important, not less. Section-level generation means a track can be altered in small pieces, blended across styles and reworked repeatedly. That creates useful creative freedom, but it also raises hard questions about attribution, provenance and whether a regenerated section is simply a new output or something closer to a derivative work.
Traditional music rights were built around recordings, compositions, sync licenses and mechanical royalties. They were not built for a system that can rewrite a verse, shift into heavy metal for eight bars, then return to a different vocal texture without breaking the song. The more flexible the tool becomes, the more pressure there is to define what was generated, what data shaped it, and who gets paid when the result is used commercially.
The timing is sensitive. Universal Music Group and TikTok renewed a licensing agreement this week that includes a commitment to remove unauthorized AI-generated music and improve attribution for artists and songwriters. That does not directly decide how ElevenLabs' model should be treated, but it shows where the industry is headed. Platforms and labels want clearer rules before AI music becomes too large to manage cleanly.
What Comes Next
For creators, Music v2 could be useful immediately. It gives them a faster way to test genres, produce demos, create background music and refine sections without starting over. For publishers and labels, it is another reason to push for better metadata, clearer opt-in systems and contracts that explain how revenue flows when AI-generated music is edited after the first output.
For ElevenLabs, the opportunity is bigger than music subscriptions. If Music v2 works reliably, the company can sell API usage, enterprise tools and licensed creative workflows to customers that already spend heavily on content. That moves ElevenLabs closer to the budgets of agencies, streaming platforms and software companies, not just individual creators experimenting with prompts.
The risk is that licensing clarity lags behind product capability. If rights holders accept the consent-first model and agree on workable accounting, Music v2 can become a serious commercial audio layer. If they do not, the same flexibility that makes the tool compelling could become the legal pressure point. The next phase of AI music will not be decided only by how good the songs sound. It will be decided by whether the industry can make the business rules as editable as the music now is.
Also read: ClickHouse's revenue surge shows why AI infrastructure is still hot • Huawei chip queen He Tingbo unveils Tau law to sidestep US sanctions • Goldman's 8,000 call says Wall Street still trusts the AI trade
