Suno launched v6 — its first models trained on licensed major-label catalogs — while UMG signed ElevenLabs to build a fan-facing remix platform, ElevenLabs shipped Music v2.5 with lossless downloads, and Spotify began flagging AI-generated artists with a new badge system.

Suno Launches v6 — the Licensed Era Arrives

Suno shipped three new models on September 9: v6 (polished flagship), v6-wild (experimental, less predictable), and v6-mini (faster, free for all users). This is the first Suno generation trained on licensed music, with Warner Music Group, BMG, and Believe all signed as partners. All previous models are being retired. New features include plain-language editing of song sections, multi-song mashups, multimodal inputs (text, audio, images, video), and a rebuilt stem separation system that regenerates stems from the model's internal representations rather than frequency-based isolation — supporting up to 100 instruments with 90%+ accuracy on vocals and drums.

The quality jump from v5 is meaningful, particularly in arrangement coherence and vocal clarity. The catch: v6-mini is all free users get, and it shows. v6 and v6-wild require Pro ($10/mo) or Premier ($30/mo). If you've been building workflows around older Suno models, start migrating — they're going away. The stem separation alone is worth evaluating as a production tool, especially the single-instrument isolation mode.

Source: Digital Music News, Bloomberg, TechCrunch, Music Ally, Variety, Rolling Stone

UMG and ElevenLabs Sign Multi-Year Deal to Build AI Music Platform

Universal Music Group and ElevenLabs announced a multi-year strategic licensing agreement on September 10. The first product is a fan-facing AI creation platform — separate from ElevenLabs' existing Music API — enabling remixes, mashups, and "personalized vocal experiences" using opted-in UMG artist catalogs. They'll also co-develop professional AI audio tools for UMG's signed artists and songwriters. This is ElevenLabs' first deal with a major label. No platform name, launch date, or pricing was disclosed.

For production professionals, the buried lede is the pro tools commitment. If ElevenLabs builds artist-facing voice and music tools with UMG's backing, that could reshape session work and vocal production workflows. For everyone else, the pattern is clear: within 48 hours of Suno's licensed launch, UMG moved to ensure it had a stake in the competing platform too. The majors aren't fighting AI music anymore — they're positioning to own it.

Source: Variety, Semafor, Engadget, Music Ally, PR Newswire

ElevenLabs Ships Music v2.5 with Lossless Downloads for All

Alongside the UMG deal, ElevenLabs released Music v2.5, an upgraded generation model with richer melodies, improved instrument layering, and more natural timbres. The bigger news for working producers: lossless downloads are now available on every plan, including Free (5 per day; Pro gets 400 per month). All plans now include commercial use rights. Downloads are blocked on tracks that reference specific artists by name.

If you've been using ElevenLabs for reference tracks or AI-generated stems, the lossless option removes the quality ceiling that made earlier versions impractical for production work. The model improvements are incremental, not generational — but the access changes are immediately useful.

Source: Futurum Group, ElevenMusic Blog

Acon Digital Acoustica 8 Brings ML Denoising to a Perpetual License

Acon Digital released Acoustica 8 on September 11, featuring machine-learning-powered DeNoise 3 and DeHum 3 modules that require no noise profiling — just apply and adjust. The update also adds ARA2 integration in every edition and immersive audio support up to 9.1.6 channels (including Dolby Atmos) in the Ultimate edition. Perpetual licenses run $149 / $299 / $449 for Standard / Premium / Ultimate, with a 25% introductory discount through October 6.

If you do audio restoration or post-production, this is worth a look. The ML denoising puts it in the same space as iZotope RX and Supertone, but with perpetual licensing instead of subscriptions. The Atmos support in Ultimate makes it a dark-horse option for immersive post workflows, especially at that price point.

Source: ProSoundWeb, KVR Audio, Sound On Sound, Production Expert

Spotify Begins Rolling Out AI Persona Badges

Spotify's AI Persona badge system, announced August 11, started appearing on artist profiles in mid-September. The badge flags artist identities that are wholly AI-generated — not human artists who use AI tools in their workflow. Labeled AI Personas are excluded from Discover Weekly, Release Radar, editorial playlists, and algorithmic recommendations by default. Artists can self-disclose, and Spotify uses human reviewers plus its own detection tools to identify profiles that meet the criteria. An appeals process is available.

If you're distributing AI-generated music on Spotify, this fundamentally changes the economics. Without algorithmic discovery, growth depends entirely on direct search and external traffic. Human artists using AI tools are explicitly exempt — the line Spotify is drawing is about whether the artist identity is real, not whether the music was AI-assisted.

Source: Spotify Newsroom, TechCrunch, Implicator.ai

SIQA Announces First AI Music Awards and Launches Provenance Registry

The Sonic Intelligence Academy announced the inaugural AI Music Awards (AIMAs) on September 10, scheduled for November 13 in Los Angeles with 25 categories. More practically useful: SIQA simultaneously launched a Verified Registry and iOS app that lets AI music creators document the origin and AI involvement of their work.

The awards are a cultural milestone, but the registry is the story for practitioners. As platforms, labels, and distributors increasingly demand transparency about how music was made, having a timestamped provenance record becomes a professional necessity — not a nice-to-have. If you're creating AI music commercially, this is worth setting up now.

Source: PR Newswire
Tool of the week
VoiceStudio

An open-source, fully local alternative to ElevenLabs that surged to 16,000+ GitHub stars this week. VoiceStudio handles voice cloning (from 3–15 seconds of audio), video dubbing, transcription, and audiobook creation across 646 languages — all running on your own hardware with zero cloud dependency. It integrates 16 TTS engines and 11 ASR engines, and includes AudioSeal imperceptible watermarking by default.

For sound designers and audio producers who need voice synthesis without per-character API costs or cloud privacy concerns, this is the most capable local option available right now. The voice cloning quality won't match ElevenLabs' best commercial models, but for prototyping, scratch tracks, and projects where data stays on your machine, it's immediately useful. AGPL-3.0 licensed — check the terms if you're building commercial products around it.

Free / Open source · Local (macOS, Linux, Windows) · github.com/debpalash/VoiceStudio
Worth watching
The walled-garden convergence

Three things happened within 48 hours this week: Suno launched v6 on licensed catalogs, UMG signed ElevenLabs, and Billboard reported that Udio has pivoted to a walled garden where AI-generated music can't leave the platform. The pattern is now unmistakable — the era of open, unrestricted AI music generation from commercial platforms is ending. What replaces it is a licensed ecosystem where major labels are partners, revenue-sharing participants, and gatekeepers of what the models can produce and where the output can go.

For independent creators and sound designers, the question is sharpening: as the big platforms license up and lock down, who builds the open tools? VoiceStudio's GitHub surge this week suggests demand is real. But for music generation specifically, open-weight alternatives like Stable Audio remain a generation behind Suno and ElevenLabs. If open AI music creation matters to your workflow, the window to advocate for it — and to support the projects building it — is narrowing.