Two Cases, One Core Conflict
In July 2026, two court decisions are due that could prove groundbreaking for how AI music generators are allowed to handle copyrighted material. In Munich, the regional court is set to rule on the lawsuit filed by German collecting society GEMA against AI company Suno, while in Boston a decision is expected in the Sony Music v. Suno case, as MusicTimes reports, within a summary-judgment proceeding. Both cases turn on the same underlying question — whether training generative music AI on copyrighted material is permissible — but they’re being litigated under different legal frameworks: German copyright law on one side, the US fair-use doctrine on the other.
Munich: GEMA v. Suno
GEMA filed its lawsuit on January 21, 2025, with Munich Regional Court I (case no. 42 O 763/25, 42nd Civil Chamber). It accuses Suno of using at least six musical works from GEMA’s repertoire — including “Forever Young,” “Atemlos,” “Mambo No. 5,” “Rasputin,” “Big in Japan,” and “Daddy Cool” — for training without a license, storing them within its models, and reproducing them in response to user prompts. The oral hearing took place on March 9, 2026; according to a report from law firm HÄRTING Rechtsanwälte, Suno contested the German court’s jurisdiction at the hearing and invoked both a fair-use argument and the text-and-data-mining exception under German copyright law.
Munich Regional Court I had originally scheduled its ruling for June 12, 2026, then postponed it “for official reasons,” according to the official press release from the Bavarian Ministry of Justice, to Friday, July 31, 2026, at 9:00 AM. The case is widely seen as a follow-on to GEMA’s win against OpenAI in November 2025, which GEMA describes as Europe’s first landmark AI ruling in the audio space.
Boston: Sony Music v. Suno
In parallel, a US case is underway that was originally brought in June 2024 by Sony Music, UMG, and Warner together — coordinated through industry group RIAA, which accused Suno of copying “decades worth of the world’s most popular sound recordings” without permission. A separate case targets provider Udio in a New York federal court. According to media reports, Warner and UMG each settled with Suno out of court (Warner in November 2025, UMG in October 2025); Sony Music is continuing to pursue the case before Chief Judge F. Dennis Saylor IV in the District of Massachusetts. Per MusicTimes, a ruling on the central fair-use question is expected there in July 2026 via summary judgment; the source gives no specific date, so this timing should be treated as unconfirmed until the decision is actually issued.
Suno’s Defense: “Fair Use” and Mathematical Patterns
In its filings, Suno does not fundamentally deny training on copyrighted material, but argues the songs it generates are not copies. According to a filing cited by Music Business Worldwide, the company states “no Suno output contains anything like a ‘sample’” from the training recordings — the AI, it argues, generates only new sounds rather than collages of existing recordings. This is Suno’s own legal position in the ongoing proceedings and has not been confirmed by any court.
The economic backdrop remains favorable for the company despite the litigation: Suno raised another $400 million in June 2026 at a $5.4 billion valuation, according to TechCrunch — right in the middle of the ongoing legal disputes.
Where This Leaves Things
Both cases test the same underlying question against different legal standards: Munich turns on the scope of Germany’s text-and-data-mining exception, while Boston will be decided under the four-factor fair-use test of US law. MusicTimes frames both pending decisions as “first substantial judicial guidance on AI music training in two major legal systems,” while cautioning that they won’t settle every open question — including compensation for independent musicians whose recordings ended up in training data without payment. Until the actual rulings land, it remains unclear how far apart German and US courts will end up in their assessment of AI music training.
