Suno is battling in court, while Udio has already made peace with labels

Suno is battling in court, while Udio has already made peace with labels

While some argue about which neural network sings better, lawyers are tackling a much duller, yet far more important question: do Suno and Udio have the right to exist in their current form at all? This July is no ordinary month for the AI music market. Two courts, in Munich and Boston, are preparing to issue rulings that will set the rules of the game for years to come. Interestingly, the two main competitors have approached this moment with diametrically opposed strategies.

Two paths: fight or negotiate

When the RIAA, on behalf of Universal, Sony, and Warner, filed lawsuits against Suno and Udio in June 2024, both services faced the same choice: negotiate with the labels or defend the position of "fair use" in court. Ultimately, their paths diverged.

Udio chose the path of settlement. Universal Music Group settled its copyright case against Udio in October 2025, and subsequently signed deals with Warner, Merlin, and the music publishers' association NMPA. But this peace came at a price: Universal and Udio are now aiming for a 2026 launch of a joint platform, which will be a "walled garden" where no AI tracks can be downloaded or published outside the site. In fact, you can generate music on Udio, but you cannot export it — and this restriction is already in effect.

Suno took a different route. The company settled the dispute only with Warner, while continuing to litigate against Universal and Sony. Despite the legal pressure, Suno surpassed 2 million paid subscribers in February with $300 million in annual revenue, and in early June, it raised $400 million in funding at a valuation of $5.4 billion — all against the backdrop of active legal proceedings. At the same time, the product is not standing still: March saw the release of version v5.5 with voice cloning, personal models, and Suno Studio — a web-based DAW for more in-depth track editing.

What will be decided this month

While users argue about whose vocals sound more natural, a much more fundamental dispute is taking place in two federal jurisdictions: whether it was ever permissible to train a model on someone else's recordings without a license.

  • July 31, Munich. The Munich court will issue a ruling on the lawsuit filed by the German collective management organization GEMA against Suno; the date was moved from June 12. The same panel of judges already ruled against OpenAI in November 2025, so the precedent does not look favorable for Suno.
  • July, Boston. A key hearing on the fair use question is being prepared in the US in the Sony v. Suno case before Chief Judge F. Dennis Saylor of the District of Massachusetts. The case has grown from the modest 560 tracks in the initial complaint to claims regarding 61,000 songs after audio fingerprinting revealed the scale of the use of protected recordings.

Why does this matter even to those who just make a jingle for an ad? Because the court's decision could determine whether fair use applies to training AI on copyrighted material — and essentially influence the entire market for generative products, not just music. Sony is fundamentally refusing to settle: as observers note, the company is "betting on a court ruling that will set a precedent by which the entire industry will live."

What this means in practice for content creators

Setting aside the legal nuances, here is what is actually changing for a copywriter, marketer, or blogger who just wants a background track or a jingle:

  • Udio is currently not suitable for distribution. Even if the sound is technically superior — Udio outputs at 48 kHz versus Suno's 44.1 kHz and features better instrument separation — you simply cannot take the track off the platform to upload it to streaming services or use it in advertising.
  • Suno's rights are clearer, but not ironclad. Paid plans provide a commercial license for created tracks, but the training of the model itself remains a subject of legal dispute — the legal landscape is still being formed, not settled.
  • Suno's toolkit is noticeably broader. Version v5.5 added voice cloning, personal models based on your own catalog, and a full-fledged workspace with stems and MIDI export — features that Udio simply lacks due to the freeze on downloads.

The practical conclusion is simple: if you need a track that can be exported and used in a project tomorrow, the bet right now is on Suno or on tools with pre-licensed data. If you care about "sound for sound's sake" — experiments and references — Udio is still good, just keep its temporary limitations in mind.

Conclusion

The AI music race has ceased to be just a question of generation quality — it is now a question of the product's legal architecture. While the courts in Munich and Boston decide the fate of the industry, a sensible strategy for those who earn money from content is not to get tied to a single tool and to keep an eye on how terms of use change.

If you want to try generating music or images, or chat with GPT, Claude, and Gemini without the hassle of access and payment, GEMERA AI has all these models gathered in one dashboard and a Telegram bot. Come in, test them, and decide for yourself which tool is right for your specific task.