Stability AI Bets on Music, With the Major Labels as Investors

Stability AI, the company best known for the Stable Diffusion image model, is reorganizing itself around music. TechCrunch reported on October 2, 2026, drawing on reporting by The Information, that Napster co-founder Sean Parker and CEO Prem Akkaraju are steering the company toward becoming a primary AI toolmaker for music professionals. The shift follows a funding round in late August in which all three major record companies took equity in the company.

The August 25 round

On August 25, 2026, Stability announced a $76 million Series B. According to The Next Web and Tech.eu, investors included Universal Music Group, Sony Music Group and Warner Music Group, along with Electronic Arts, AMD Ventures and Pacific Alliance Ventures. Returning backers included Coatue, Greycroft, Kadmos Capital, former Google CEO Eric Schmidt and Parker himself. Stability said the round brings total funding raised since Akkaraju became CEO in June 2024 to $232 million. No valuation was disclosed.

Universal and Warner already had strategic partnerships with Stability, signed in October and November 2025 respectively. "Our investors set us up for long-term success because they provide not only capital, but expertise, credibility, and direct connection to artists," Akkaraju said in the announcement, as quoted by Tech.eu. Stability described its own position as building on "fully licensed data" and argued that "a licensed platform has to be better, not merely cleaner."

TechCrunch reported that, as part of the deal, the three music companies also licensed their catalogs for training. The companies' public announcements focused on the investment itself, and terms of any catalog licenses have not been published.

What Stability is building

In May 2026, Stability released Stable Audio 3.0, a family of four models. Music Business Worldwide reported that the company said all four were trained on licensed data from the AudioSparx production library (806,284 audio files) and Creative Commons recordings from Freesound. Three of the models are open-weight and downloadable; the largest is available only through an API. The mid-sized model can generate tracks up to six minutes and 20 seconds long.

According to TechCrunch, Stability has since released three more audio models and AI music-editing software. The tools generate full instrumental tracks or short snippets from text prompts, and Parker says an upcoming update will let users hum a melody or beatbox a drum pattern to guide the output. TechCrunch also reported that Parker contrasted the approach with Napster's, conceding that his earlier company asked for forgiveness instead of permission, and saying Stability is seeking permission first.

The Next Web noted two other details: Stability moved its headquarters from London to Los Angeles and closed its UK office, and the company continues to face copyright claims in other parts of its business.

How this differs from Suno and Udio

The contrast with the consumer song generators is in the product shape. Suno and Udio built their audiences by generating complete songs with vocals from a text prompt, then faced lawsuits from the labels, followed by settlements with some of them. Stability is approaching the labels as investors before scaling a music product, and its tools so far center on instrumentals, editing and sound design for professionals. That positions it closer to production software than to a replacement for songwriting.

For the majors, the stake gives them a financial interest in at least one AI music company whose growth does not depend on unlicensed training, alongside their licensing deals with Suno, Udio, ElevenLabs and others. It also diversifies their bets at a time when Suno's licensed v6 models are drawing user complaints and fresh litigation from Sony and Universal.

Our analysis: open questions

Several things remain unclear from the public record. The terms of the catalog licenses have not been disclosed, so it is not known how revenue will reach the artists and songwriters whose work trains the models, or whether individual artists can opt out. It is also not yet known whether label-licensed training data will produce better results than Stable Audio 3.0's production-library data, or whether professional users will adopt the tools at scale.

What is clear is the direction of travel. In 2024 the major labels' main relationship with AI music companies was litigation. In 2026 it increasingly includes equity, licensing and co-development, and the companies they back describe their products as tools for musicians. Whether that promise holds will show up in product decisions over the next year.

What it means for independent artists

If AI tools for instrumentals and editing become standard studio software, the scarce input becomes what those tools do not generate: a strong song and a recognizable voice. Independent writers should track how these licenses pay creators, since the answer will set expectations across the market. Starchild's approach is a working example on a small scale: the catalog is built on human-written songs, every song carries an AI-transparency label, and the style models used to produce alternate versions are licensed, with revenue shared with the artists behind them.

Sources

• TechCrunch: Sean Parker is rebuilding Stability AI around music
• The Next Web: Stability AI raises $76mn from Universal, Warner, Sony Music and EA
• Tech.eu: Stability AI raises $76M, with Warner and Sony investing
• Music Business Worldwide: From the major labels' $76M Stability AI bet to Universal and Sony's Suno 'stream ripping' claim
• Music Business Worldwide: Stability AI launches new audio models that can generate 6-minute music tracks