Apple Music has asked a judge to stop Suno from taking testimony about parts of its streaming business, escalating a discovery dispute that puts AI-music royalties, platform data and confidential label negotiations in the same legal spotlight.

The request arises from the continuing copyright litigation brought by Universal Music Group and Sony Music against Suno. Those music companies allege that Suno trained its AI tools on their recordings without permission and that the resulting tracks compete with their music. Suno has maintained that its conduct is protected by fair use. Warner Music Group was also originally a plaintiff but reached a settlement with Suno in 2025; the Universal and Sony matter remains active, with evidence gathering expected to continue through mid-2027.

At issue now is not simply whether Apple Music has material from Suno on its service. Apple says it has already provided extensive documents, including material responsive to six document subpoenas from Suno and the labels. That production included information about how it identifies AI music and how much Suno material is present on the platform. Apple’s objection is aimed at a demand for testimony it says would expose commercially sensitive information that goes well beyond what Suno needs for its defense.

Why a streaming platform is part of the case

Apple Music is not described as a defendant in the Universal and Sony lawsuit against Suno. It has instead become a third party from which the litigants seek evidence. That distinction matters. A company outside the central dispute can be compelled to supply relevant evidence, but it can also contest demands it considers excessive, irrelevant or improperly invasive of its business.

The information Suno sought reportedly extends to minimum payments, the treatment of AI songs in royalty arrangements, streaming fraud and Apple’s licensing negotiations with Universal and Sony. One reported line of inquiry asked whether Apple Music could remain in business without the major labels’ catalogues.

That is a vast question with obvious commercial sensitivity. It touches the negotiating leverage of streaming services and music companies, the importance of catalogues in a subscription platform’s offering, and potentially the economics behind licensing relationships. Apple’s position is that this terrain is not necessary for resolving the core claims against Suno.

Apple also argues that the plaintiffs themselves are better positioned to provide relevant information. The company pointed to more than 90 emails and over 10 discussions conducted across more than a year as the parties worked out what Apple would provide. In other words, Apple’s argument is not that it has supplied no evidence; it is that document production should be enough and a further examination would intrude on protected business details.

Discovery is the pretrial process in which parties obtain evidence. It can include documents, written questions and testimony under oath. Discovery is designed to ensure that each side can test the other’s claims with evidence rather than surprise, but it is often where the hardest arguments occur because businesses may hold sensitive material that is relevant only at the margins.

A subpoena is a formal demand for evidence from a person or organization. In this dispute, Apple says it responded to six document subpoenas. Document subpoenas can seek records such as internal communications, metrics, agreements or policies. A request for testimony goes a step further by seeking answers from a company representative.

Testimony in this setting generally means sworn questioning. It can be useful when documents require explanation or when a party believes it needs to learn how a business made decisions. But it also creates greater risk of disclosing confidential operational details, particularly where the questioning reaches licensing negotiations, payment structures or strategic dependency on catalogues.

A protective order is a court mechanism that can limit the use or disclosure of sensitive information. Apple has asked either for the testimony demand to be cancelled or for an order protecting its information. The first route would prevent the requested testimony; the second could allow some discovery while placing boundaries around what is disclosed and how it may be handled.

None of those procedural questions decides whether Suno’s training was lawful. They do, however, shape what evidence each side gets before a court addresses the larger copyright conflict.

Royalties are the practical pressure point

The fight over Apple’s testimony highlights why AI music litigation is not only a debate about technology or abstract copyright doctrine. It is also about the route from a play to a payment.

Royalties are payments tied to the use of music. The supplied details do not establish how any particular Apple Music agreement treats Suno tracks, what its minimum payments are, or whether AI songs receive treatment different from other recordings. Those are precisely the details Apple says Suno is trying to explore. But the questions themselves show why the issue matters: if AI-generated tracks are distributed alongside conventional recordings, decisions about licensing, detection, fraud controls and compensation can affect who receives streaming revenue.

The labels contend that Suno-made tracks compete with their recordings. If that argument is accepted, the alleged harm is not limited to the copying claimed during training. It can also concern listener attention and streams once AI tracks reach platforms. Suno’s fair-use defense, meanwhile, places the legal focus on whether its use of music in training can be excused under copyright law. The two positions frame the technology’s impact very differently.

Apple’s request emphasizes a separate concern: a platform may have valuable evidence without becoming an open book on every aspect of its commercial relationships. Questions about the labels’ importance to Apple Music, for example, could reveal more about bargaining positions and private business arrangements than about whether Suno’s training is fair use.

AI identification and fraud controls are part of the picture

Apple has already provided information concerning how it finds AI music and the amount of Suno material on the service. That is notable because the ability to identify AI music is relevant to the real-world administration of a streaming catalogue. A platform needs some way to understand what material is being uploaded or distributed if it is to assess rights questions, respond to claims and manage its catalogues.

Suno’s requests also covered streaming fraud. In simple terms, streaming fraud concerns activity that artificially generates plays or otherwise manipulates systems connected to music consumption and payment. The available information does not allege that Suno committed streaming fraud, nor does it describe Apple’s specific anti-fraud practices. It only establishes that the subject formed part of the requested inquiry.

That boundary is important. AI music, fraud and royalties can overlap in a single discovery request without becoming the same allegation. The case described here remains a copyright dispute over training and alleged market competition. Apple’s application concerns the scope of evidence Suno can obtain from the streaming service.

The wider legal pressure on AI music companies also includes claims centered on producer tags. These are short signature sounds associated with producers and often heard in rap music. Their presence can be commercially and creatively significant because they identify a producer’s work or brand to listeners.

Music publisher Reservoir has separately accused Suno of generating tracks that included producer tags. Its lawsuit sought $157.5 million in connection with 1,050 works. Reservoir alleged that Suno trained its systems using songs separated into vocals and backing tracks, as well as cover versions. Those allegations are distinct from the Apple discovery dispute and from the ongoing Universal and Sony litigation.

Reservoir also filed a separate copyright case against Anthropic concerning lyrics, seeking $127.95 million tied to 853 works. The parallel cases demonstrate that arguments about generative systems can involve different kinds of music material: complete recordings, components of recordings, lyrical works and recognizable production signatures. Each claim has its own factual record and legal issues, so none should be treated as a finding in the others.

For producers, the stakes can be especially direct when a recognizable tag appears in generated music. For labels and recording artists, the focus may include the recordings used in training and whether AI releases substitute for licensed catalogues. For streaming services, the difficult work involves handling music that reaches listeners while protecting confidential commercial relationships and responding to rights disputes.

What happens next

Apple has asked the court to quash the testimony demand or impose protections for its confidential information. The motion was listed for December 1, leaving the next decision with a judge. A ruling could determine whether Suno receives further answers from Apple and, if so, under what restrictions.

It would not end the broader case. Universal and Sony’s claims against Suno are still in the evidence-gathering stage, and that schedule is set to run through mid-2027. As a result, the Apple dispute is best understood as one important procedural battle within a much longer contest over permission, training data, AI-made tracks and the value of music streams.

The issue also reaches beyond one service or one AI company. Platforms need workable ways to identify AI material and administer catalogues. Rights holders want to protect recordings and the revenue attached to them. AI developers seek legal room to build and operate their tools. Where those interests collide, the most consequential answers may not come only from the eventual fair-use ruling, but also from the records and business information courts decide may be examined along the way.

Gaming platforms are confronting a related quality-and-AI governance question, though in a different market: Valve has addressed AI and low-quality submissions around Steam Next Fest. The common thread is not that music and games face identical legal rules; it is that distribution platforms increasingly sit between automated creation tools, rights concerns and audiences trying to find legitimate work.