Reservoir Sues Suno and Anthropic Over AI Music Copyright, Seeking Nearly $338 Million
Independent music publisher Reservoir Media has filed two major copyright infringement lawsuits against artificial intelligence companies Suno and Anthropic, accusing both businesses of using copyrighted music without permission to train their AI models.
The lawsuits, filed on 21 October 2026 in the US District Court for the Northern District of California, concern more than 1,000 musical compositions and hundreds of copyrighted song lyrics.
Reservoir is seeking a combined maximum of approximately $338 million in statutory damages across the two cases, alongside injunctions and other remedies.
The legal action adds to a growing number of disputes between music rights holders and AI developers, raising important questions about whether copyrighted music can be used to train generative AI without licensing agreements.
Why is Reservoir suing Suno and Anthropic?
Reservoir alleges that both companies copied copyrighted material from its publishing catalogue without obtaining the necessary permissions.
However, the lawsuits involve two different types of AI technology.
Suno operates a generative AI music platform that allows users to create songs from text prompts. Anthropic develops Claude, a general-purpose AI assistant capable of generating written content.
Reservoir argues that both businesses have benefited commercially from copyrighted musical works without properly compensating the rights holders.
The company has identified compositions associated with artists including Joni Mitchell, Johnny Cash, Madonna and Evanescence.
Among the songs referenced in the legal filings are:
- Both Sides Now – Joni Mitchell
- Ring of Fire – Johnny Cash
- Papa Don’t Preach – Madonna
- Bring Me to Life – Evanescence
- Big Yellow Taxi – Joni Mitchell
Reservoir says it fully owns or exclusively controls the relevant US copyrights in the works identified in its complaints.
The allegations have not yet been proven in court.
Reservoir’s Suno lawsuit involves 1,050 copyrighted compositions
The lawsuit against Suno identifies 1,050 musical compositions that Reservoir alleges were copied without authorisation.
According to the complaint, Suno obtained music from online sources to create the training datasets used by its generative AI models.
Reservoir alleges that the company used automated scraping methods to collect content from platforms including YouTube Music, Deezer and Genius.
The lawsuit also references publicly reported information from leaked Suno source code, which Reservoir argues provides evidence of how the company’s training datasets were assembled.
Crucially, the dispute isn’t limited to whether Suno copied existing recordings.
Reservoir also challenges the platform’s ability to generate new music that allegedly reproduces or derives from protected musical compositions.
The publisher argues that Suno’s restrictions on imitating established artists can be circumvented through particular prompts.
These claims could become important when courts consider the distinction between AI systems learning musical characteristics and generating outputs that infringe protected creative expression.
Reservoir is seeking approximately $186.4 million in statutory damages under the copyright and Digital Millennium Copyright Act claims outlined in the complaint.
Anthropic faces a separate lawsuit over 853 songs
Reservoir’s lawsuit against Anthropic focuses primarily on copyrighted song lyrics.
The publisher alleges that lyrics from 853 compositions were copied and used to train Claude without permission.
It also alleges that the AI assistant can reproduce protected lyrics, or generate unauthorised derivative versions, when responding to user requests.
The Anthropic complaint includes works such as Joni Mitchell’s Big Yellow Taxi and Both Sides Now, alongside songs associated with Madonna and Evanescence.
Reservoir is seeking approximately $151.4 million in statutory damages across the claims detailed in the complaint.
This isn’t Anthropic’s first dispute involving music copyright.
The company has already faced legal action from publishers including Universal Music Publishing Group, Concord, ABKCO, BMG and others over the alleged unauthorised use of song lyrics and compositions.
Anthropic has disputed Reservoir’s allegations and indicated that it intends to defend the case.
Why are the lawsuits potentially worth $338 million?
US copyright law allows rights holders to seek statutory damages for qualifying infringements.
For proven wilful copyright infringement, the potential statutory award can reach $150,000 per work, although actual awards depend on the facts and legal findings.
Reservoir’s Suno complaint identifies 1,050 works, producing a claimed maximum copyright infringement figure of $157.5 million.
The Anthropic complaint identifies 853 works, producing an equivalent figure of $127.95 million.
Both lawsuits also include claims under the Digital Millennium Copyright Act relating to allegedly circumventing technological access controls and removing copyright management information.
Those additional claims bring the combined figures pleaded across both lawsuits to approximately $337.8 million.
Importantly, these amounts represent damages requested by Reservoir, not money that has been awarded or agreed in a settlement.
Reservoir is also seeking court orders that could require the removal or destruction of unauthorised copies held in training datasets, internal repositories and model infrastructure.
Such remedies, if granted, could have significant implications for how AI developers maintain their training data.
Suno’s existing music licensing agreements make the case particularly interesting
One of the most significant arguments in Reservoir’s lawsuit concerns Suno’s existing licensing relationships with music companies.
Suno has reached agreements with music businesses including Warner Music Group, BMG and Believe.
Reservoir references these agreements as evidence that a commercial market already exists for licensing copyrighted music to AI developers.
That raises an important question.
If AI companies can negotiate permission to use copyrighted music today, what happens to music that was allegedly copied and used before those agreements existed?
Reservoir argues that the existence of licensing agreements supports its position that rights holders should be compensated for the use of their work.
However, a company entering into licensing agreements doesn’t automatically establish that its previous training practices were unlawful.
That question remains for the courts to determine.
There is also a distinction between licensing sound recordings and licensing musical compositions.
A recording and the underlying composition can have different copyright owners. A licence covering one set of rights won’t necessarily authorise the use of the other.
For generative AI developers, this creates an additional challenge when negotiating rights across large music catalogues.
Can AI companies legally train their models on copyrighted music?
This is one of the biggest unresolved legal questions surrounding generative AI.
AI developers have argued in copyright disputes that training models on existing material can constitute transformative fair use under US law.
The argument broadly centres on whether training an AI system to identify patterns and generate new material represents a legally permissible use of existing copyrighted works.
Music publishers and record labels have challenged this interpretation, particularly where copyrighted material has allegedly been copied without permission, obtained through unauthorised methods or reproduced in AI-generated outputs.
Reservoir’s complaints raise several related questions.
Does copying music into a training dataset constitute infringement? Does the way that material was obtained affect the legal analysis? And when does an AI-generated output cross the line from musical influence into copyright infringement?
These issues aren’t identical, and courts may reach different conclusions depending on the evidence and particular uses involved.
The Reservoir lawsuits could help shape future interpretations of copyright law, but they don’t yet establish a legal precedent determining whether all AI training requires a licence.
What do the lawsuits mean for independent musicians?
For independent artists and songwriters, the most important issue is whether the growing commercial value of AI music technology will translate into appropriate recognition and compensation for music creators.
Generative AI platforms can produce music at enormous scale, potentially creating new opportunities for musicians while also introducing additional competition and copyright concerns.
There are several potential implications.
Music publishing rights are becoming increasingly important
Much of the public discussion about AI music focuses on recordings and artists’ voices.
However, Reservoir’s lawsuits demonstrate the importance of the underlying composition, including melodies, lyrics and songwriting rights.
Independent musicians should understand how their publishing rights are administered and who has authority to license their compositions.
AI music licensing could create new revenue opportunities
As AI companies negotiate agreements with rights holders, licensed music datasets could develop into an additional revenue market.
This doesn’t guarantee payments to individual musicians, and the financial terms of many agreements remain confidential.
Nevertheless, music rights are becoming an important commercial consideration in the development of generative AI products.
Independent rights holders need representation
Major record labels and publishers have substantial resources to negotiate licensing arrangements and pursue infringement claims.
Smaller music businesses and independent songwriters may have less direct negotiating power.
This makes transparent licensing frameworks, accurate rights ownership data and effective rights administration particularly important.
Any future AI licensing market will need to address how independent creators can participate alongside larger catalogue owners.
The music industry’s AI licensing battle is far from over
Reservoir’s lawsuits against Suno and Anthropic are part of a much broader legal battle over the relationship between copyright and artificial intelligence.
Some AI music developers have started negotiating licensing agreements with major music companies, while others remain involved in substantial copyright disputes.
The Independent Music Publishers International Forum has publicly supported Reservoir’s legal action and called for fair licensing arrangements that recognise the value of songwriters’ and composers’ work.
Ultimately, the debate isn’t simply about whether artificial intelligence should have a place in music.
It’s about establishing the rights, permissions and commercial arrangements that allow new technology to develop alongside the creators whose work may be used to build it.
The outcomes of these cases could influence the future economics of AI music generation, publishing rights and licensing agreements for years to come.
For independent musicians, the central question remains: as AI companies build increasingly valuable businesses around music, how will the people who create that music participate in the value generated?