Udio signs Kobalt while fighting indie artists in court, OpenBMB drops open-source creative TTS, Hatsune Miku gets a modern VOCALOID6 voice, and a game studio reverses its AI soundtrack decision after backlash.
Udio inks Kobalt — fourth licensing deal before its "new platform" launch
Announced April 9, Udio's partnership with Kobalt follows deals with UMG, Warner, and Merlin, and will feed Udio's forthcoming relaunched subscription service with authorized songwriter catalog. The product is pitched as a remix/cover/new-song tool where voices and compositions of opted-in Kobalt artists become usable source material, with credit and payout. For anyone doing sync, ad, or content work, this is the clearest signal yet that AI music is splitting into a licensed tier (Udio, Spotify/labels, probably Suno) and an unlicensed one.
Udio simultaneously moves to dismiss independent artists' class action
Same week as the Kobalt deal, Udio is aggressively pushing to dismiss the October 2025 class action brought by indie artists over unlicensed training data, arguing Illinois has no jurisdiction. Read together: majors get licensing partnerships; indies get motions to dismiss. If you make the kind of music that gets trained on without a major behind you, the message is stark.
OpenBMB drops VoxCPM2 — open, tokenizer-free TTS for creative sound design
Released April 11: an open multilingual TTS model that skips the traditional tokenizer step, which the team argues frees output from "phoneme-locked" flatness. Explicitly positioned for creative audio content and voice cloning — not just narration. If you've been boxed in by ElevenLabs' API pricing or wanted self-hosted expressive vocalizations (whispers, screams, non-speech vocals) for game/film work, this is the first credible open option in months. Free, self-hosted, research-license.
Hatsune Miku VOCALOID6 library drops April 14
Yamaha is releasing a new Hatsune Miku voicebank for VOCALOID6 on April 14, with Japanese and English at launch and Chinese coming later. This matters beyond Vocaloid fandom because VOCALOID6's new synthesis engine has closed a lot of the expressivity gap with Synthesizer V — and Miku is the most-used synthetic voice in the world. If you write to picture or make pop demos, a modern Miku voice is suddenly a much more serious production tool than the meme version most people remember.
Saving Country Music publishes explicit AI policy for 2026
Independent country/Americana outlet Saving Country Music formalized a stance against covering AI-generated music on its platform, citing the flood of fake artists on streaming services. It's a single publication, but it's the first major genre-specific outlet to publish a bright-line editorial rule — expect more to follow, especially in roots/country/folk where authenticity is core to the listener relationship.
Generative game audio hits a backlash moment — Chronicles Medieval reverses course
Upcoming RPG Chronicles Medieval publicly walked back its decision to use generative AI for its soundtrack after community pushback, re-labeling all assets as "100% handcrafted." Combined with industry data showing ~63% of game devs worried about AI copyright, this is the first high-visibility case of a studio reversing an AI soundtrack decision — useful leverage for composers pitching game work right now.
Worth revisiting because the AI vocal synth landscape just keeps getting more serious, and Synth V 2 Pro remains the best practitioner tool most producers still haven't tried. Unlike ElevenLabs (prompt-driven, black-box) or Suno/Udio (full-song), this is a DAW-adjacent editor where you draw MIDI, type lyrics, and shape vowels, breath, timbre, and pitch with surgeon-level control. Blind-test rated as "human-level natural" by 50 musicians.
For songwriters who need a pro vocal demo at 2am, or composers who need a proxy singer with no studio booking, it's the one AI voice tool that feels like an instrument rather than a prompt box.
The Udio story this week is the clearest example yet: the same company that just signed its fourth major-catalog licensing deal is simultaneously fighting to dismiss lawsuits from independent artists on technical jurisdictional grounds. Spotify's October 2025 label-partnership framework, Suno's Warner deal, Udio's four licensing deals — all flow to rights-holders with the lawyers to force a deal. Independent artists, whose catalogs made up a meaningful chunk of the scraped training data, are getting procedural defenses.
Two rulings to watch: the GEMA v. Suno decision on June 12 in Germany and the US fair-use ruling expected this summer. Either could force retroactive licensing for the long tail. If you're an independent artist, the window to register with your collection society and be counted in any future settlement pool is now — not after a class is certified.