Suno Loses German Copyright Case, Must License Music for AI Training

A German court ruled against Suno on July 31, finding that the AI music generation company violated copyright law by training its models on protected recordings without obtaining licenses. The decision requires AI firms operating in Germany to secure explicit licensing agreements before using copyrighted music in model development, marking one of the most consequential intellectual property rulings for the AI audio sector to date.

The court's reasoning cuts directly against the fair use arguments AI companies have deployed in similar litigation across the United States and Europe. German copyright law does not recognize a broad fair use exemption, and the ruling confirms that commercial AI training does not qualify for any of the narrow exceptions available under German statute. For Suno, a company whose core product depends on having absorbed vast libraries of recorded music, the practical implications are immediate: continued operation in Germany now requires licensing deals with rights holders, adding a structural cost that did not exist before today.

Music rights holders have pursued this outcome methodically. European collecting societies and record labels have filed parallel actions against multiple AI firms across Germany, France, and other EU member states over the past two years, building a litigation record that courts are now drawing on. The German decision fits a pattern in which European courts have sided with rights holders in disputes over AI training data, including earlier rulings against text-to-image generators that found scraping copyrighted visual works for commercial model training to be infringement.

The European AI Act, which entered phased application in 2025, adds a regulatory layer on top of the copyright question. The Act requires providers of general-purpose AI models to publish summaries of training data and to comply with EU copyright law, including opt-out mechanisms for rights holders. The Suno ruling reinforces that compliance posture: courts are treating the AI Act's copyright provisions not as aspirational guidance but as enforceable obligations with teeth.

AI developers counter that licensing at scale is operationally complex. Music rights are fragmented across labels, publishers, and individual artists, and there is no single clearinghouse that can grant a training license covering the global recorded music catalog. Suno and its peers argue that requiring per-work or per-catalog licenses before training would effectively make large-scale AI music models impossible to build legally, concentrating the market among incumbents with existing label relationships. Critics of the ruling also note that Germany's decision does not automatically bind courts in other EU member states, meaning the licensing obligation may apply unevenly across the bloc until either the Court of Justice of the European Union weighs in or the European Commission issues harmonizing guidance.

None of those arguments changed the outcome. The ruling stands, and its immediate effect is to establish Germany as a jurisdiction where AI music companies face enforceable licensing obligations. Other AI firms active in the German market, including those developing models trained on text, images, or video, should read the decision carefully. The court's logic does not appear narrowly tailored to music; it applies the same underlying copyright principles that have governed other creative works for decades. If appellate courts or the CJEU affirm the reasoning, the licensing requirement could extend well beyond the audio sector.

For music rights holders, the German decision is a meaningful enforcement win after years of watching AI companies train on creative catalogs without compensation. Whether it translates into a workable licensing framework or simply pushes AI music development into friendlier jurisdictions will depend on how quickly the industry and collecting societies can negotiate terms. The legal pressure is now clearly established. The commercial negotiation is next.