A profile of UMG CEO Lucian Grainge, who helped record labels rake in billions of dollars from streaming and now hopes to do the same with generative AI
John Seabrook / New Yorker : X: @carolineonline_ and @drorpoleg . Forums: Hacker News See also Mediagazer X: Caroline Butler / @carolineonline_ : A bit of a puff piece for Lucien Grainge but an absolutely fascinating read on a fearless approach to AI in music and how necessary controlled experimentation is. https://www.newyorker.com/... Dror Poleg / @drorpoleg : “A decade ago... major labels were responsible for about ninety per cent of the content on streaming platforms; today, their share is roughly ten per cent.” 😮 @jmseabrook's exploration of AI vs. music: https://www.newyorker.com/... Forums: Hacker News : The music industry's A.I. experiments See also Mediagazer
Context & Ripple Effects
UMG’s streaming-era playbook is the backdrop: rights holders turned platform distribution into a major revenue source even as major labels’ share of streaming content reportedly fell sharply. The earlier negotiations over AI voice tools on YouTube show that AI music was already becoming a licensing question, not just a technology question.
Grainge’s position matters because UMG can seek to make access to recordings, voices, and catalog rights part of the commercial terms for generative tools. The later emergence of prompt-driven song generation with human-like vocals underscores why labels are focused on control at the point of creation.
First-order effects
- UMG’s AI posture shifts toward controlled experimentation aimed at licensing and monetizing generative uses of music rather than treating the technology solely as a threat.
- Artists and repertoire represented by UMG become central commercial inputs in any AI products the label chooses to authorize, increasing the importance of consent and rights management.
Second-order effects
- AI music developers and distribution platforms face stronger incentives to pursue label agreements if they want access to commercially valuable catalogs, voices, or artist participation.
- Other major labels are pressured to define comparable licensing positions, while unlicensed generators face a more adversarial rights environment, as reflected in the later RIAA litigation against Suno and Udio.
Third-order effects
- If licensing becomes the dominant route to market, generative music may develop a two-track structure: governed products with cleared rights and open tools carrying greater legal and commercial uncertainty.
- The strategic contest shifts from whether AI can generate music to who controls training inputs, identity rights, provenance, and revenue allocation across an expanding volume of synthetic tracks.
The trend: Generative AI is pushing music rights holders to adapt the streaming-era licensing model to synthetic creation, with governance becoming a competitive asset.