Amid legal battles, Suno says it will start watermarking songs
TechCrunch Ivan Mehta
Suno's adding watermarks to AI songs and tightening download rules to stop fake tracks flooding streaming sites. It's basically damage control while labels, courts, and a data breach lawsuit pile up against the company.
Suno spent much of 2025 getting sued, and now it's trying to look like the responsible adult in the room. The AI music startup announced today it's rolling out audio watermarking and fingerprinting to tag songs made on its platform, plus new limits on downloads meant to stop people from mass-uploading AI tracks to Spotify and other streaming services and quietly collecting royalties. CEO Mikey Shulman framed it as a principled stand for transparency in a blog post, but the timing says otherwise.
The company also inked a deal with lyrics database Musixmatch to use its Sentinel system for catching copyright violations, though Suno wouldn't say whether it's building its own watermarking tech or borrowing something like Google's SynthID. Community guidelines got a rewrite too, now explicitly banning deceptive audio passed off as real and cloning someone's voice without permission. Shulman was careful to note the tools won't judge whether a song is any good or human enough, just whether it's honest about its origins. He's punting the disclosure question to artists and platforms, which is a convenient way to avoid taking a firm stance on anything.
And there's plenty of pressure to explain the sudden interest in guardrails. Suno is currently being sued by Universal Music Group and Sony Music Group in a case backed by the RIAA, and a German court just ruled against the company on behalf of licensing body GEMA, finding it violated copyright law. Then there's the November data breach, first reported by 404 Media, which exposed that Suno had scraped YouTube, Deezer, and Genius to train its models in the first place. Have I Been Pwned later confirmed the breach hit 55 million users, and Massachusetts residents have since filed a class action accusing Suno of skipping security basics to chase growth.
Suno raised $400 million in a Series D round back in June, so money isn't the problem. Credibility is. Watermarks and download limits might slow down the copycat problem on streaming platforms, but they do nothing to address the underlying accusation that Suno trained its models on scraped, unlicensed content without asking anyone first. That's the fight actually playing out in courtrooms right now, and no amount of fingerprinting fixes it.
My take
Watermarking is the easy, cheap gesture companies make right before a judge tells them the hard part actually matters. Suno wants credit for policing what happens after the music gets made while dodging the much bigger question of what it fed its models to make that music possible in the first place. Expect more of this pattern across AI companies facing lawsuits: announce a safety feature, hope it buys goodwill, keep the actual training data fight buried in legal filings nobody reads.
Read more about this at: TechCrunch
