Stability AI raises $76M from major record labels and AMD to fuel AI audio tools

Stability AI raised $76 million in a Series B round from Sony Music Group, Universal Music Group, Warner Music Group, and AMD Ventures, signaling that the biggest names in entertainment see strategic value in AI-powered audio and music production tools. For AI builders, the round funds continued open-source releases like Stable Audio 3.0 and deeper partnerships that could shape how audio generation models are integrated into creative workflows.

Major record labels and AMD back a $76M round

The consortium includes Sony Music Group, Universal Music Group, and Warner Music Group, along with AMD Ventures, Electronic Arts, and other investors. Stability AI says it will use the proceeds to develop additional products for creative professionals and grow its applied research and professional services teams. The round brings Stability AI's total funding to $232 million under new leadership.

Stable Audio 3.0 brings single-step diffusion-transformer generation

Stability AI released Stable Audio 3.0 in May, a diffusion-transformer-based audio generation suite. It includes three open-source models and a paid API that can generate audio clips up to six minutes and 20 seconds long. The architecture uses a single-step generation process: it starts with random noise and replaces it with coherent sounds in one pass instead of many. That is faster than traditional multi-step diffusion.

The model represents audio in a compressed latent space using a semantic-acoustic autoencoder, which Stability AI claims speeds up inference without reducing output quality. The autoencoder incorporates transformer elements to preserve detail. For builders, this means lower latency for audio generation APIs and the ability to run longer clips without proportional compute cost.

Partnerships with music labels and Electronic Arts point to commercial audio tools

Two of the three label backers, Universal Music Group and Warner Music Group, are already partnered with Stability AI to develop AI-powered music production tools. Electronic Arts, also an investor, previously worked with Stability AI on image generation for game design and plans to use the company's models to speed up in-house workflows. These relationships suggest that the next wave of Stability AI products will be tightly integrated into existing creative pipelines rather than sold as standalone generators.

What AI builders should watch

The round signals that open-source audio models are becoming commercially viable in regulated industries like music. Builders integrating audio generation into their products should watch for:

• API availability and pricing: Stable Audio 3.0 already has a paid API accessible via cloud. Pricing details were not disclosed in the announcement, but the label partnerships may influence how copyright and attribution are handled in generated audio.
• Open-source model quality: The three open-source models from Stable Audio 3.0 are available for self-hosting, giving developers control over latency and data privacy. Benchmark comparisons against other audio generation models were not provided in the announcement.
• Partnership-driven features: Tools developed with Universal and Warner may offer genre-specific fine-tuning, style controls, or licensing hooks that general-purpose audio APIs lack.

The round is a bet that AI audio tools will not displace creators but become part of their toolset, analogous to how plugin ecosystems evolved in music production. Builders should treat the label involvement as a strong signal that commercial audio generation will require clear provenance and licensing infrastructure, not just high-quality output.