AI music startup Suno has rebuilt its core technology on a licensing foundation. On Wednesday, the company introduced Suno v6, a new family of music generation models it says was developed using licensed data from labels and distributors including Warner Music Group, BMG and Believe.
Crucially, Suno says v6 was not trained on the data behind its earlier models — the same data that triggered a wave of litigation from the recorded music industry. The company plans to retire its older models entirely.
What’s actually launching
Suno is shipping three variants, each aimed at a different use case:
- Suno v6 — the base model, available to paying users, positioned as reliable and steerable for controlled outputs.
- Suno v6 wild — an experimental model, also for paying users, built for ideation and deliberately unexpected results.
- Suno v6 mini — a faster version available to all users, including free accounts.
The feature set is where things get interesting for anyone producing branded audio. Suno v6 allows users to edit part of a song via a prompt or by changing a word in the lyrics. It accepts text, images or video as reference inputs for generating tracks. And it can isolate an instrument from a sample so a new beat can be built around it.
The company also flagged a remix capability on the roadmap, dependent on artists opting into a new programme it is building with labels that would allow Suno to use their songs in AI-generated features.
Licensing as a revenue pitch, not just a legal shield
Suno is framing the shift as a commercial opportunity rather than a compliance exercise. Chief product officer Jack Brody told TechCrunch that the music ecosystem and Suno’s partners are "always looking for ways to create more revenue opportunities" for rights holders and artists, and said a significant part of the release is about creating additional revenue streams.
That framing matters. The generative AI content debate has largely been argued in courtrooms; Suno is attempting to move it into a rights-and-royalties conversation that labels recognise and can price.
The legal picture is far from settled
Suno settled with Warner Music Group last year and struck a deal with BMG last month. But it still faces suits from Sony and Universal Music Group, from artists including Jason Isbell, and from users alleging the company deprioritised security in pursuit of profit. The v6 announcement also landed a day after Suno acknowledged training its models on YouTube videos.
Alongside the model launch, the company has been tightening controls: it said last month it would watermark songs generated on its platform, and it recently introduced download limits tied to account tier. Brody said Suno can help manage ecosystem problems such as streaming fraud and low-intent mass uploads to distributors, but argued that governing what content appears on platforms ultimately falls to distributors and the platforms themselves.
Suno has raised more than $819 million to date, per PitchBook data.
Why marketers should care
Generative audio has quietly become a working tool for social edits, performance creative, retail radio, podcast beds and regional campaign adaptations. The blocker has rarely been quality — it has been indemnity. Brand safety teams and legal departments have been reluctant to sign off on assets built with models of uncertain provenance.
A licensed training set changes that risk calculus, at least partially. But it does not eliminate it while major-label litigation is live. A sensible checklist before you put AI-generated audio into a paid campaign:
- Confirm which model version produced the asset, and whether that version’s training data is licensed.
- Check the commercial-use terms attached to your subscription tier, not just the platform’s marketing claims.
- Understand watermarking — how it appears, and whether it affects broadcast or platform delivery specs.
- Keep a provenance log of prompts, references and outputs for every asset that goes live.
- Ask your agency partners the same questions before they hand over a cutdown.
The broader signal for martech buyers: licensing is becoming a product feature. Expect image, video and voice tools to compete on data provenance the way they currently compete on output quality.
Source: TechCrunch
