Reservoir Media is the independent publisher behind songs by Joni Mitchell, Johnny Cash and Madonna. On Wednesday (Oct. 21) it filed two federal copyright cases, Billboard reports: one against Suno, the other against Anthropic. This post covers what the Reservoir Suno lawsuit claims, what the separate Anthropic case is about, the full timeline, and what it changes if you put AI-generated music in your videos, streams or podcasts.
Same court, same day, same lawyer. It’s the legal version of ordering two coffees at once because you already know how the afternoon is going to go.
The Reservoir Suno lawsuit in numbers
The Reservoir lawsuit against Suno was filed in the US District Court for the Northern District of California, Music Business Worldwide reports. The plaintiffs are Reservoir Media Management, Inc., its UK subsidiary Reservoir Media Management (UK) Limited, and Shapiro, Bernstein & Co. Limited, which the UK entity wholly owns.
So the plaintiffs list reads like corporate nesting dolls, each one holding a copyright. Somewhere there is an org chart with arrows that loop back on themselves. Anyway.
Music Ally reports that Reservoir’s Suno lawsuit focuses on 1,050 compositions it owns or exclusively controls in the US. A composition is the song itself (melody and lyrics). Not a recording. The songs named across the two complaints include Joni Mitchell’s Both Sides Now, Johnny Cash’s Ring of Fire, Madonna’s Papa Don’t Preach and Evanescence’s Bring Me to Life.
Then there’s the money. In the Suno lawsuit, Reservoir asks for actual damages and Suno’s profits or, alternatively, statutory damages of up to $150,000 per work willfully infringed. Statutory damages are a range set by law, so a rights holder doesn’t have to prove exactly what it lost. The complaint puts the top of that range at $157.5 million. Digital Music News adds up both suits and gets to nearly $340 million.
The DMCA claims and a ten-dollar sentence
Copyright isn’t the only charge. The Reservoir lawsuit also accuses Suno of circumventing access controls and removing copyright-management information. In plain words: getting around the locks on sites that host music, then stripping out the metadata that says who owns a song. Those two counts come with their own price tags, $2.625 million and $26.25 million. MBW adds that only Suno is named as a defendant, although the complaint alleges Suno used Bright Data’s tools and services to get around those controls.
Reservoir’s attorney also has a way with a sentence. Busch writes that Suno built a platform where anyone able to pay $10 a month can make complete songs based entirely on these illicit copies. That’s Richard S. Busch of Adams & Reese, known for the Blurred Lines case, and he represents Reservoir in both complaints.
What happened so far, from the first Suno lawsuit to Reservoir
This didn’t start on Wednesday. Here’s the timeline, as the coverage tells it:
- June 2024: Universal Music Group and Sony Music Entertainment first sue Suno.
- November 2025: Warner Music Group settles. BMG and Believe later sign licensing deals, MBW notes.
- August 2026: Round Hill Music sues Suno and Anthropic, each suit with 500 compositions as “a prioritized representative bellwether”, MBW reports.
- September 18, 2026: UMG and Sony sue Suno again. This time: 60,202 recordings.
- October 21, 2026: Reservoir files its Suno lawsuit (1,050 compositions) and its Anthropic lawsuit (lyrics of 853 compositions).
- October 22, 2026: the International Music Publishers Forum says it welcomes and supports Reservoir’s action. Anthropic says it disagrees with the allegations and will defend the case vigorously. Suno has not commented.
A confession. At the start of the year we thought one post could keep track of the AI music cases. It couldn’t. Even our Suno v6 explainer ended up needing a section about lawsuits.
The Anthropic case is about lyrics, not audio
The second complaint is a different animal. Music Ally reports it focuses on the lyrics of 853 compositions, which Reservoir says were used to train Claude. It also says the chatbot produces copies and derivative works of those lyrics when users ask, and publishers have strong opinions about a machine that can recite their catalogue on request.
That’s why this counts as a two-front attack. One target makes music from prompts. The other is a general chatbot. Publishers now treat both as the same problem, because lyrics are half of what they own.
If this sounds familiar, it should. A group of publishers has been fighting Anthropic over lyrics for a while, and we covered the fair use hearing in that older case. Reservoir brings a fresh catalogue to the same argument.
Suno’s licensing deals are now evidence in the Reservoir case
This is the part we’d frame. The Reservoir lawsuit against Suno turns the company’s own agreements with Warner Music Group, BMG and Believe into evidence. According to Digital Music News, Reservoir argues that by signing them Suno “all but admitted that there is a functioning market for licensing music and sound recordings for training data.”
Why does that matter? AI companies usually defend training with fair use, the legal doctrine that allows some uses of copyrighted work without permission. When courts weigh fair use, they ask if the use harms a market for the work. If a licensing market for training data plainly exists (because Suno is paying into it), that defence gets harder to argue.
The complaint describes the strategy less politely, per Billboard: “infringe now, make billions by being the first mover, and try to settle on the cheap later”.
What we think: licensing later has a cost nobody put in the pitch deck. Every deal that calms one rights holder becomes an exhibit for the next one. It’s like apologising to one neighbour in writing, then finding your letter stapled to three more complaints.
What the Reservoir Suno lawsuit means for creators using AI music
First, the obvious point. These are allegations, and no court has ruled on them. Reservoir’s Suno lawsuit names the company, not the people who typed prompts into it.
One request still deserves attention. Reservoir wants an injunction and the impoundment or destruction of unauthorized copies of its works, MBW reports, including copies held in model weights. Model weights are the numbers a model learned in training, so basically the model itself. Would a court ever order that? Nobody knows. Still, if your channel depends on one generator, its legal future is now part of your workflow.
Then the boring, true point: a tool’s licence is only as solid as the tool’s own rights to what it learned from. A paid plan tells you what the company promises. It doesn’t tell you what a publisher with 1,050 songs thinks of that promise.
One number from our side. 25.9% of our copyright checker results found a rights holder’s claim on the uploaded track, slightly down from 27.3% (Mubert’s own data, last 28 days). That’s about one in four. Lawsuits decide who pays for training data over the next few years. Claims decide if your video earns anything this week.
What we think, and what to do this week
The pattern is clear now. The majors went first, then independent publishers piled in: Round Hill in August, Reservoir in October, both going after Suno and Anthropic in pairs. IMPF says AI companies must now come to the table and negotiate fair licences. So expect more deals and more complaints, often in the same week.
While the Reservoir and Suno lawsuits are pending, a few habits will save you a headache:
- Keep a record of every AI track you publish: the tool, the plan, the date and the prompt.
- Reread the tool’s licence for commercial use now, not on the day a claim arrives.
- Keep artist names and real lyrics out of your prompts.
- Run finished uploads through a copyright checker before they go public.
- Prefer tools that say where their training audio comes from and if the musicians behind it get paid.
Two complaints in one Wednesday, three plaintiffs, 1,050 songs and 853 sets of lyrics. Somewhere in Northern California, a court clerk is ordering a third coffee.
Who we are
