SAN FRANCISCO, California — Independent music company Reservoir Media has filed two federal lawsuits accusing Suno and Anthropic of copying copyrighted music and lyrics without permission to train artificial intelligence systems.

The complaints were filed October 21 in the U.S. District Court for the Northern District of California. One case alleges Suno infringed 1,050 Reservoir-controlled compositions; the other alleges Anthropic copied lyrics from 853 compositions for its Claude models and generated infringing outputs.

Reservoir’s catalog claims include works associated with Joni Mitchell, Johnny Cash, Madonna, Miles Davis and Evanescence. The allegations have not been proven, and neither court has ruled on liability or remedies.

Reservoir seeks actual damages and profits or, alternatively, statutory damages. Its pleaded maximums across the copyright and Digital Millennium Copyright Act counts total about $186.4 million against Suno and $151.4 million against Anthropic — roughly $337.8 million combined.

The requested relief goes beyond money. Both complaints ask for injunctions and the impoundment or destruction of unauthorized copies of Reservoir works, including alleged copies retained in training datasets, internal repositories, model weights and server infrastructure.

Those demands could have practical consequences for AI music services and rights licensing if a court ultimately accepts Reservoir’s theories. For now, they remain requests by the plaintiff, not court orders.

Reservoir’s Suno complaint is case 3:26-cv-12305. Its Anthropic complaint is case 5:26-cv-12308. Music Business Worldwide first detailed the filings and published Reservoir Chief Executive Golnar Khosrowshahi’s statement that the company intends to protect its songwriters, artists, estates and partners.

Radio News Now had not identified public responses from Suno or Anthropic to the new Reservoir complaints at publication time. Both companies have disputed infringement theories in other pending music-related cases.

Radio News Now question: Will courts treat model weights as copies subject to destruction, or will the disputes push AI companies and independent publishers toward catalog-wide licensing?