Universal Music Group has signed a multi-year licensing agreement with ElevenLabs, a company specializing in voice models and generative audio. The deal involves creating a new consumer-facing AI-powered music platform and jointly developing audio products for artists and songwriters. For ElevenLabs, this marks its first agreement with a major record label; for UMG, it is another step toward building a music AI market based on permissions, rights management, and new commercial uses of its catalog.
The platform is still in development and will be separate from the music products ElevenLabs already offers. The stated goal is to allow fans to creatively engage with music from participating artists and songwriters: features cited by the companies include remixes, mashups, new interpretations of tracks, and personalized vocal experiences. Neither a launch date nor the business model for users, rights holders, and participants has been disclosed.
It is precisely this lack of details that characterizes the current phase of the deal. UMG and ElevenLabs have announced an industrial framework, not a market-ready product: it remains to be clarified which catalogs will be included, what permissions will be required from artists, how much control they will retain over the outputs, and how compensation will be calculated. In generative music, the decisive challenge is not proving that a model can create or transform audio, but establishing who is allowed to do so, under what terms, and with what return for the creators of the original work.
A deal that brings AI into the artist-fan relationship
Until now, the discussion around AI-generated music has primarily focused on two fronts: models trained on copyrighted catalogs and services that generate tracks from text prompts. The alliance between UMG and ElevenLabs shifts the focus toward fan engagement. The promise is not just generating a song from scratch, but offering regulated ways to rework authorized music and voices, turning the catalog into an interactive environment.
For labels, this could open up an extension of the activities that already fuel the digital music business: alternative versions, personalized content, engagement campaigns, and formats that extend the commercial lifespan of a release. For artists and songwriters, in theory, the platform can offer a way to tap into these creative languages without ceding control to external services that lack a direct relationship with rights holders.
However, the outcome will depend on the concrete rules. A fan-made remix, for instance, involves recording, composition, performance, and sometimes elements of an artist's vocal identity. A product designed for the global market will have to navigate territorial differences in licensing, collecting societies, and rights availability. It will also need to make clear to users the boundary between officially authorized content and a creation that can only circulate within the platform.
UMG presented the agreement as part of an ecosystem of partners aimed at expanding the creative and commercial possibilities of human artistry. Mati Staniszewski, co-founder and CEO of ElevenLabs, emphasized the combination of the company's models and UMG's expertise in rights management, along with a commitment to ensuring fair compensation for artists and songwriters. These are central principles in the debate, but the financial terms of the deal and the revenue-sharing mechanisms have not been made public.
Why ElevenLabs is entering a new category
The move comes as ElevenLabs commands a financial scale rarely seen in the generative audio sector. In February, the company raised $500 million in a Series D round that valued it at $11 billion. That figure more than doubles the $5.4 billion valuation reached in June by Suno following its $400 million round. The comparison does not directly measure product quality or the ability to impact the music market, but it signals how much capital and investor confidence have flowed into AI audio.
Founded in 2022 by Staniszewski and CTO Piotr Dąbkowski, ElevenLabs was built and expanded primarily in the synthetic voice, audio infrastructure, and conversational agent sectors. Headquartered between London and New York, it reports that its products are used by employees at two-thirds of Fortune 500 companies. In May, it announced that it had surpassed $500 million in annual recurring revenue, up from the $350 million reported at the end of 2025. This enterprise base sets the company apart from platforms built solely around song generation.
Music is nonetheless not new territory for ElevenLabs. In August 2025, it launched ElevenMusic, an application for generating and editing original tracks, alongside deals with Merlin and publisher Kobalt. It later added an agreement with Believe, and in April stated that more than 4,000 independent and emerging artists were using the service. Alongside the app is Music API, which makes the company's music models available to businesses and developers.
The difference is that the project with UMG will not be folded into these services: it will be a distinct offering, tied to a major label and its rights network. This is a significant choice from both a reputational and operational standpoint. Separating the authorized platform from general-purpose tools can help delineate catalogs, permitted uses, and liabilities, preventing the public from confusing creations made within a licensed environment with content generated elsewhere.
The platforms' race toward licensing
The announcement comes one day after the launch of Suno's v6 suite. The company stated that it developed those models with Warner Music Group, BMG, and Believe; its Chief Product Officer Jack Brody clarified that the training data does not include recordings from UMG. The two developments highlight an industry rapidly moving away from an abstract standoff between the music business and AI to enter a phase of selective deals, where each catalog can be included, excluded, or made available under different terms.
This does not mean tensions have been resolved. Licensing deals with a few major corporations do not equate to universal consent from creators, nor do they eliminate concerns regarding training, voice cloning, attribution, and competition with human labor. But they do reshape the bargaining landscape: platforms looking to attract top-tier catalogs must prove they can provide traceability, controls, and compensation, rather than merely technology capable of delivering convincing results.
For UMG, the bet is that generative interaction can become a product category distinct from traditional streaming, without devaluing the catalog. For ElevenLabs, the test will be turning its audio expertise into a music service that simultaneously satisfies fans, artists, publishers, labels, and rights holders. Upcoming announcements will need to address whether the platform will be rolled out in select markets, whether it will require subscriptions or purchases, and, above all, what opt-in and control tools it will offer to the artists involved.
The value of the deal therefore lies not just in ElevenLabs's billion-dollar valuation or UMG's global scale. It lies in the attempt to establish an operational framework for the creative use of music with AI: not an undifferentiated library to generate from freely, but a system where access, transformations, and monetization must be negotiated. If it succeeds, it could shape how other platforms present their proposals to majors and indies alike.
