Stability AI's Series B and label backers: what it means for audio
Stability released Stable Audio 3.0 on 5/20/26 and raised a Series B on 8/25/26 with EA, Sony, UMG and WMG. A dated timeline and what to verify for video work.

Stability AI announced a Series B on August 25, 2026 with Electronic Arts, Sony, UMG and WMG named among the participants, and its news page describes Stable Audio 3.0 as trained on fully licensed data. For anyone making video with generated music, the practical meaning is that licensing provenance is now a headline feature of an audio model, not an afterthought.
The dated timeline
These dates come from Stability's news and updates page as read on October 3, 2026.
| Date | Event |
|---|---|
| 5/20/26 | Stable Audio 3.0 released, trained on fully licensed data |
| 8/18/26 | DAW plugin and StableAudio.com interface announced |
| 8/25/26 | Series B with EA, Sony, UMG and WMG |
What the news does and does not tell you
A licensed training set and label investors reduce one category of risk, but they do not replace reading the terms for the specific model or plan you use. The page does not state what each plan permits for commercial video, how attribution works, or what happens to downloaded files if a plan lapses. Those are the details a producer needs before shipping an ad.
Treat the announcement as a reason to ask better questions, not as a clearance.
- Does the plan you will use allow paid advertising and client work?
- Are generated files covered after you cancel?
- Is there an API, or only a plugin and a web interface?
- Who is the licensor of record if a platform asks for proof?
Keeping a paper trail either way
Whatever model you choose, store the model id, the date, the prompt and the output URL next to each finished video. On Sume, a music job returns an audio artifact on media.sume.com, and the job record keeps the request. The Music Router docs state that job.request.routed_model names the catalog engine that ran, which gives you the engine id for your file without guessing.
A one-line log per track is enough: date, job id, routed model, where it was used. If a client later asks where a soundtrack came from, you can answer in a minute.
curl https://api.sume.com/v1/jobs/job_123/result \
-H "Authorization: Bearer $SUME_API_KEY"
# keep: job id, date, routed model, artifact URL, where the track shippedTakeaway
The Stability news is a signal that audio vendors are competing on licensing. For video teams the useful response is to compare each vendor's commercial-use terms in writing and to keep a record of every track you ship.
Sources
Related posts
More in Media tools
- Stabilize shaky video by API: Sume has no deshake, so do this
Sume's video filter refuses deshake and vidstab as unknown filters. Prove it with the free check, stabilize upstream, then use Sume for trim, crop and captions.
- Strip the English voiceover before localizing: audio-detach is 1 cent
Pull the audio track from a finished ad with Sume audio-detach at $0.01 per job to reuse, review or replace it before a 12-market localization pass.
- Temu detail video up to 180 s: stitch 30 s clips with Timeline
Temu detail video allows up to 180 s and 300 MB at 720p or higher. Sume clips top out at 30 s, so six clips stitched with Timeline cover the full length.
- TikTok trending API: browse without a query is flag-gated
Sume trending search needs a query on production. Browse with no query, limit up to 100 and relevance floors are dev-only until a flag is set. The differences.
Written by Sume