Udio

In April 2024, five former Google DeepMind researchers emerged from stealth with a product that the music world had not seen before: a text-to-music generator capable of producing fully mastered songs — with realistic vocals, layered instrumentation, and original lyrics — in under 40 seconds. Rolling Stone tested a pre-release beta and within a single day generated two tracks so sonically close to Tom Petty's voice that the Petty estate was contacted for comment. The product was Udio. The team — David Ding, Yaroslav Ganin, Charlie Nash, Conor Durkan, and Andrew Sanchez, all veterans of DeepMind's Lyria and Imagen projects — had spent four months building it after leaving Google. Andreessen Horowitz led the seed round. will.i.am and Common wrote checks. Instagram co-founder Mike Krieger and Oriol Vinyals — head of Gemini at Google — joined as angels. The message from the music and tech establishment was clear: this was the real thing.

The legal chapter that followed was inevitable. In June 2024, the Recording Industry Association of America filed suit on behalf of Universal Music Group, Warner Music Group, Sony Music, and others, alleging that Udio (and Suno) had trained their models on copyrighted recordings at industrial scale without permission or payment — seeking up to $150,000 per infringed work, a figure that in theory could reach hundreds of billions of dollars. Rather than fight the labels into bankruptcy court, Udio chose a different path. In October 2025, it settled with Universal Music Group, agreeing to rebuild on exclusively licensed music. In November, it settled with Warner. In January 2026, it struck a deal with Merlin, the global rights agency representing thousands of independent labels. The terms were consistent: Udio would transition to a platform trained only on authorized, licensed content, with participating artists and songwriters credited and compensated when their sonic DNA appears in a generated track. The vision David Ding had articulated from day one — a platform that benefits artists, not just users — was finally being built on contractual ground rather than legal risk.

The transition has not been frictionless. Downloads were temporarily restricted during the licensing negotiations, prompting a user revolt that forced CEO Andrew Sanchez to reopen downloads for 48 hours. The creative community that had grown up around Udio's previously unrestricted generation capabilities — artists using the platform to create without-limits music, content creators building entire YouTube channels around AI-generated tracks, producers using it to test arrangements — now operates within a walled garden of licensed sources. The debates about whether this makes Udio more sustainable or simply more limited will outlast the lawsuits. What is not in debate: Udio is now building toward the only thing that makes a consumer AI music platform defensible long-term — a legal and creative infrastructure that labels, rights holders, artists, and regulators can all live with. At $8M ARR with $10M raised, it is the smaller of the two major AI music platforms, competing directly with Suno's $300M ARR and $2.45B valuation. But it is the one that chose to work with the music industry rather than fight it, and in 2026, that bet may prove to be the more durable one.