Sony and UMG Sue Suno Over 'Poisoned' AI Model v6

Key points

• Sony Music and Universal Music Group filed a 45-page lawsuit in the U.S. District Court for the District of Massachusetts on September 18, 2026.
• The suit alleges Suno’s new v6 models, launched with partners Warner Music Group, BMG, and Believe, still infringe on copyrighted works.
• Labels claim that training v6 on outputs from prior models 'launders' previous infringement, calling it the 'fruit of the same poisoned tree.'
• The complaint cites 60,202 specific recordings allegedly infringed, claiming this is only a small portion of the total violations.
• Suno responded that its claims are 'fundamentally flawed,' stating v6 was trained on licensed content and user interactions, not unlicensed data.

Background

This legal action follows a June 2024 lawsuit by Sony, Universal, and Warner Music Group against Suno for training models on unlicensed tracks. In late 2026, Suno pivoted by signing licensing deals with Warner, BMG, and Believe to launch v6, while Sony and Universal continued litigation. Recent months have seen increased legal friction, including a $1 billion suit from Round Hill Music and class-action claims from artists like Jason Isbell regarding AI-generated likenesses.

How outlets are covering it

Variety and Engadget emphasize the legal argument that v6 inherits the copyright violations of its predecessors, noting that Suno’s admission of using user outputs from earlier models undermines its claim of clean training data. Billboard highlights the industry’s fragmented response, noting that while BMG and Warner have signed deals with opt-in clauses for artists, many managers and lawyers remain skeptical about how these agreements protect individual creators. MusicRadar focuses on user backlash, reporting that some subscribers are cancelling over v6’s perceived lack of quality, though this is distinct from the legal dispute. Suno maintains that its new models are compliant and that the labels' claims are legally and factually incorrect.

Why it matters

The lawsuit tests whether AI models can be deemed compliant if they are partially trained on outputs from previously infringing systems. It highlights the tension between major labels seeking strict copyright enforcement and others pursuing licensing deals, potentially setting a precedent for how AI training data provenance is judged in court.

What to watch

The case will proceed in federal court, where judges must determine if training on outputs from infringing models constitutes a new violation. Suno may face statutory damages and attorney fee orders if found liable. Meanwhile, the industry will watch how other labels and artists navigate opt-in versus opt-out clauses in AI licensing deals.