AI Companies Are About to Relearn Napster's Most Expensive Lesson

What is a song worth to a machine that already learned from it?

On July 31, 2026, the 42nd Civil Chamber of the Munich Regional Court delivered the first concrete answer, ruling that Suno infringed copyrights held by GEMA, the German music rights society, by training its music models on protected works without a license. The court ordered Suno to disclose the revenue tied to that use and to pay damages. Suno can appeal, so nothing about the decision is final.

It is GEMA's second victory over an AI company in under a year, following a Munich ruling that sided with the society against OpenAI in November 2025 over song lyrics reproduced in ChatGPT. That repetition matters more than either verdict alone, because it shows the institutional response to AI training is no longer theoretical: the institutions that collect for songwriters have found a repeatable path to the courthouse and used it twice. The court also added a point with extraterritorial reach, holding that training conducted in the US does not shield a company once its output is stored and reproduced in Europe.

I watched the first version of this movie from inside the building. Early in my career I worked at Liquid Audio and then at Napster, during the music industry's first digital reckoning. I remember the feeling in those hallways: smart engineers building a product people loved, and a slow certainty settling in that the law would arrive before the business model did. It did.

Strip away the legal terminology and the Munich ruling reduces to something plain: training data is an input, inputs have owners, and owners get paid. For years the industry has argued that a model learning from a work is fundamentally different from copying it, and courts in several jurisdictions are still weighing the same argument. In Munich it lost.

The question underneath the headline is older than AI: who captures the value a new technology creates, and who pays for the relearning it demands? Every song Suno trained on had a writer, someone who sat with a guitar in a small room and finished a chorus, and GEMA exists to collect for precisely those people. A disclosure order reads as an accounting exercise inside a courtroom. To a working songwriter, it is the first time anyone has been made to show the receipt.

The music business met its first digital disruption with a sequence I still recognize in boardrooms today: ignore it, then shame the people using it, then regulate it. Napster opened in 1999 and a court shut it down in 2001, but the audience demand never disappeared. By 2003 Apple's iTunes Store was selling songs for 99 cents to the same listeners the industry had been suing. The market the labels feared became the market that carried them, once the input finally had a price.

Suno is a different company in a different decade, and none of this is an invitation to pile on; the response is the part that rhymes. What Munich did was pull the licensing reckoning forward, and that is a gift even to the companies paying for it, because every year a technology operates without a price is a year of value nobody can plan around.

What changes for AI companies that train on music

For an enterprise leader, this lands as a supply chain story rather than a music story. Your models learn from something, somebody owns that something, and if your legal team cannot name the sources of your training data today, you are carrying a liability you have not seen priced. Napster sat in that position once, and the bill arrived regardless.

Courts settle who pays. Leaders decide who learns.

This is where judgment earns its keep, because the work that lasts sits Above the Algorithm: deciding what your organization feeds a model, and attaching a name to that decision. A person makes that call, in a meeting, on a Tuesday.

And if you own neither the model nor the budget, the question still belongs to you, because the material you paste into a tool at your desk came from somewhere too. Ask where, and ask out loud. That question is cheap now and expensive later.

I keep the wider perspective on how rulings like this reach enterprise teams here: AI governance.

Zoom out and the consequences run wider than one company, because the value AI generates has to land somewhere. The skills required to work with licensed, governed AI have a half-life that keeps running whether or not anyone appeals. There is a precedent for closing exactly this gap. Film and television faced the same clearance problem decades ago and answered it by building a dedicated profession. Music supervisors and rights-clearance teams verify every second of a soundtrack against its rightful owners before broadcast, with a professional guild and a career ladder behind them. Curriculum half-life usually measures how quickly a course goes stale after it is written. Munich has just created the mirror problem, because the clearance work for training data now exists inside every AI company, and the course that should feed it has not been written.

Which brings me to the job this ruling created and nobody has posted yet. Somebody inside every company that trains on creative work now has to read the provenance record for each dataset before the next training run, and clear or reject the sources that arrive without a license. Somebody has to own the revenue accounting a court can order opened, the way Munich just ordered Suno's. Call it training-data clearance, the same opening the labels filled after 2003 when they built licensing teams for the digital market they had tried to sue away. Today that work is nobody's job description and everybody's exposure.

I do not know how the appeal will be decided, and neither does anyone else. But the question travels well, and it costs nothing to ask in your own building: who clears what our models learn from, and who signs? Take it into your next leadership meeting and watch how quickly the room answers, because the answer you receive is the real state of your AI program. And if you see it differently, I am easy to find.

Sources: Munich Regional Court (Landgericht München I), 42nd Civil Chamber, ruling of July 31, 2026, GEMA v. Suno · GEMA press statements, July 31, 2026 · Music Week, "GEMA wins court ruling on breach of copyright by AI music firm Suno."

Here is what makes Alex a credible voice on this topic: he worked at Napster and Liquid Audio during the music industry's first licensing reckoning, and has spent the years since helping boards decide what their companies feed a model before a court decides it for them.

Bring Alex to your next board session on AI governance. Book a call →