October 3, 2026
Music labels invest in music generation: Stability AI raises $76 million with their backing
Sean Parker is rebuilding Stability AI around music, with music labels helping fund the company.
On Aug 25, Stability AI announced a $76 million Series B round. The music investors included the rights holders themselves.
Participants in the round included: - Sony Music Group; - Universal Music Group; - Warner Music Group; - Electronic Arts.
Under Prem Akkaraju's leadership, the company has raised $232 million. Parker, who previously clashed with the music industry, is now developing music generation with support from the labels.
Longer than earlier clips. Stable Audio Open Small generated 11 seconds of audio. In the Stable Audio 3.0 lineup, introduced on May 20, Small creates recordings up to 120 seconds long, while Medium reaches 6 minutes 20 seconds. For a game or app, that is enough for background music rather than a short sound effect.
Where to work with audio
Editor and plugin. The StableAudio.com web editor and a plugin for music software have been available in beta since Aug 18. In the editor, you can modify results with text instructions, combine multiple tracks and export a mix. It also lets you adjust volume and stereo placement and apply effects.
The plugin is available for macOS in AU and VST3 formats and runs on Apple Silicon and Intel. It matches generation to the project's tempo and creates tracks up to 6 minutes long. In Ableton, open it through PlugIns on an Instrument track; in Logic, through the Instrument slot.
An existing recording can be reworked using audio-to-audio. In the official example, the model turns beatbox.wav into a lo-fi hip-hop beat at noise level 0.5. The recommended range of 0.3–0.9 controls how much the recording changes.
What is available for your own setup
Weights for Small SFX, Small and Medium are available on Hugging Face. Large runs through the Stability AI API and is available for enterprise deployment on your own infrastructure. As of Oct 2, the official README recommends installing with `uv sync` and adding the Gradio interface with `uv sync --extra ui`.
Small runs on CPU. Medium requires a CUDA GPU and Flash Attention 2. Medium's training dataset contains 1,278,902 recordings: 806,284 licensed from AudioSparx and 472,618 sourced from Freesound.
Original source: [Sean Parker is rebuilding Stability AI around music — TechCrunch, Oct 2, 2026](https://techcrunch.com/2026/10/02/sean-parker-is-rebuilding-stability-ai-around-music).
Plugin: [Download Stable Audio](https://stableaudio.com/plugin).
The Community License allows commercial use of Stable Audio 3.0 outputs, while organizations with annual revenue above $1 million need an Enterprise License.
