Sony Music Entertainment has filed a sweeping copyright lawsuit against Udio, the AI music generation platform, accusing it of infringing more than 30,000 songs in its catalogue. The Verge reported the filing, which was made in a New York court and covers recordings ranging from Elvis Presley's Hound Dog to Beyoncé's Say My Name and Harry Styles' As It Was.
The sheer scale of that song list is the first thing worth sitting with. Thirty thousand tracks is not a list assembled to make a narrow legal point. It is a statement of intent, a signal from one of the world's largest music rights holders that it views AI music generation not as a peripheral nuisance but as an existential commercial threat. Sony Music's catalogue is among the most valuable in the entertainment industry, spanning decades of popular recording and representing enormous royalty streams. The decision to enumerate thousands of specific titles rather than make a broader, more abstract claim appears designed to demonstrate the breadth of alleged harm and to put weight behind arguments that training on copyrighted material constitutes infringement at an industrial scale.
This lawsuit does not arrive in a vacuum. The major record labels have been engaged in an escalating legal campaign against AI audio companies throughout the past year or so. Universal Music Group, Sony Music Entertainment and Warner Music Group collectively filed suit against both Udio and Suno, another AI music generation startup, in earlier actions. Those cases have been working through the courts, and the new filing against Udio specifically suggests that Sony is either unsatisfied with the pace or scope of existing proceedings, or is building a parallel case intended to strengthen its overall litigation position. The record industry's legal strategy here appears to mirror what publishers and visual artists have done against image-generation AI companies, though the music cases carry particular weight given how directly the outputs of these tools compete with the industry's core product.
Udio, for its part, is one of a small group of startups that emerged in the past couple of years offering users the ability to generate full songs, with vocals and instrumentation, from text prompts. The technical achievement is significant, and the user experience is compelling enough to have attracted substantial attention. But the business model rests on a foundation that the major labels are now aggressively contesting in court. The central legal question is whether training a generative AI model on copyrighted recordings constitutes infringement under existing law. The AI companies have broadly argued that training constitutes fair use, a transformative process that does not substitute for the original works. The labels, unsurprisingly, reject this framing entirely.
The consequences of how courts resolve these questions will extend well beyond Sony and Udio. A ruling that training on copyrighted material without licensing constitutes infringement would reshape the economics of the entire generative AI sector, not only for music but potentially for every medium. The startup community would face a stark choice between licensing deals that could prove prohibitively expensive and rebuilding models on licensed or synthetic data. The licensing path is one the major labels clearly prefer, since it would position them as gatekeepers to one of the most valuable data inputs in the AI industry, generating revenue streams that the current unlicensed training environment entirely bypasses.
For Udio specifically, the litigation creates immediate and serious operational uncertainty. Defending against a case of this scale is expensive, and the reputational shadow over the platform will affect both its ability to attract users and its conversations with investors. Smaller AI music companies watching these cases are likely reading the litigation calendar closely, since whatever precedents emerge will determine whether their own models face similar exposure.
The choice of New York as a venue is also worth noting. Federal courts in the Southern District of New York have substantial experience with complex intellectual property cases, and the music industry in particular has litigated major disputes there over many decades. Sony's legal team will be familiar with the terrain.
What to watch for next is relatively clear. The immediate question is whether Udio mounts an aggressive fair use defense, seeks a settlement, or attempts to negotiate some form of licensing arrangement that might resolve or narrow the dispute. The parallel proceedings against other AI audio companies will also provide signals about how receptive courts are to the labels' arguments, and any early rulings on motions to dismiss or summary judgment will be closely read by every party in the AI and entertainment industries. Perhaps most consequentially, Congress and regulators have been circling questions about AI and copyright for some time without producing definitive guidance. A significant judicial ruling in any of these active cases could accelerate pressure on lawmakers to act, or alternatively give one side enough leverage to reshape the landscape before legislators find the will to do so.
