She hit play and a stranger’s voice filled the studio. I watched a label executive stop mid-note, hand over the demo and frown. You could feel the room split—excitement on one side, alarm on the other.
I follow this story because you should know what it means when the gatekeepers start to sign checks. I’ll show you how Suno’s v6 works, why the labels made a deal, and what it does — and doesn’t — solve for artists and listeners.
In a cramped negotiation room last winter, lawyers passed printed playlists across the table.
That room produced Suno’s new model, v6, and a legal framework few expected. Warner Music Group, BMG and Believe agreed to let Suno train on portions of their catalogs and use artist likenesses when musicians opt in. The settlement with Warner followed a wave of lawsuits and public blowback after reports showed earlier models trained on music scraped from the open web and YouTube audio — Suno has admitted using YouTube material in training.
Jack Brody, Suno’s chief product officer, told Bloomberg the company will pay labels and publishers for the copyright material it trains on, with the labels deciding how to slice royalties among artists and songwriters. The payment mechanics are still being negotiated, but the legal handshake gives Suno something few AI music projects had before: cooperation from major rights holders. This arrangement feels like a treaty signed inside a drumbeat — messy, practical, and oddly musical.
On a late-night test I asked v6 to graft a chorus onto a demo and watched it work fast.
v6 accepts prompts that reference songs, photos, videos, journal entries or vague “vibes” and can edit parts of an existing track, assemble mashups, or generate new material informed by licensed catalogs. Suno says the model was trained on content from Warner, BMG and Believe, which lets it reference styles and timbres with permission rather than relying on unlicensed scraping.
How will royalties be split when AI trains on label catalogs?
Short answer: the labels will decide. Longer answer: Suno will pay labels and publishers for the rights it uses; those entities then determine payouts to songwriters and performers. I’ve spoken with publishing lawyers who warn that line-item audits and backend deals can take months. You should expect opaque settlements at first, then more transparent splits only if artists insist and if labels build clear reporting tools.
In courtrooms and private talks, journalists kept replaying the same clip of scraped audio.
The legal record is messy. Universal Music Group, Sony Music and Warner once sued Suno (and also targeted Udio); Warner later settled with Suno and Udio. Universal reached a separate settlement with Udio while Sony and Universal’s case against Suno continues. Public leaks and a 404 Media report exposed models trained on thousands of hours of internet-sourced music, igniting suits from artists like Jason Isbell and a broader industry backlash.
The Warner settlement included the explicit option for artists and songwriters to license their recordings and likenesses to Suno. That’s a pivot from adversarial litigation to a transactional model: labels monetize their catalogs and AI firms gain licensed data. Whether that model preserves creative control for artists — or merely monetizes their past work — is the fight you’ll be watching next.
At a small club, a session guitarist asked me if machines will take his next gig.
That fear is widespread. Some artists — Dr. Dre among them — have signaled openness to AI tools as production assistants. Others worry AI will replace session players, reduce demand for human songwriters, and produce sterile output that floods streaming services. Critics also argue quality suffers when generative systems mimic without lived experience, producing what many call soulless music.
Can AI-generated songs legally chart on streaming services like Spotify?
They already can. AI-created tracks have appeared on platforms and even climbed charts. The difference now is whether streams come with licensed rights and royalty flows attached. Spotify is building tools with Universal and Merlin to let users create AI remixes and covers from participating artists — a paid add-on that Alex Norström described as the first legal way for listeners to join the AI remix wave. If that product scales, you’ll see AI songs chart with official licensing rather than as contested uploads.
At the label level, executives treat v6 as both risk management and opportunity.
Labels see v6 as a way to monetize catalog value they would otherwise police. Artists and advocates see it as a lever to force payments from an industry that once treated sampling and covers as gray zones. For you as a creator, the immediate implications are practical: you can ask whether to opt in, demand contract language about royalties and attribution, and push for transparent reporting on uses of your voice or compositions.
Suno’s move is not a clean fix. The model reduces one legal vector — unlicensed training on major catalogs — but it doesn’t remove moral questions or the economic pressure on working musicians. The promise of licensed AI is that it turns unauthorized copying into paid access; the peril is that the market simply absorbs faster, cheaper content and squeezes living creators.
I’ll keep watching contracts, courtroom filings and the first artist payouts that follow v6’s rollout. You should too, because what happens next will shape who gets paid when an algorithm borrows your chorus — and who gets blamed when a hit sounds hollow. Is this the moment the music industry buys its way into AI legitimacy, or the moment artists demand real control and pay for their voices?