What happened:
Google and Apple just moved AI music from a niche toy to a default setting. Google’s Gemini now generates 30-second tracks with Lyria 3 (Google DeepMind’s music model), adds SynthID watermarks to mark AI-generated audio, blocks blatant copying, and even auto-creates cover art with Nano Banana (Google’s image model). It’s rolling out on desktop today and on mobile in a few days, for users 18+ in eight languages. Paid plans get higher generation limits. Read more on the Google blog.
Rights and limits:
Google says users have a right to use the tracks they generate. At the same time, it checks outputs against existing content to avoid IP or privacy violations. Free users can make about 10 tracks per day, while paid tiers range from roughly 20 to 100 tracks per day depending on plan. Translation: fun, but not unlimited. More detail at latimes.com.
Apple’s move:
Apple took a lighter route but still mainstream. Playlist Playground, an Apple Music feature powered by Apple Intelligence (Apple’s system-wide AI), turns a text prompt into a 25-song playlist with cover art and a short description. It’s in the iOS 26.4 developer beta now, with a broader rollout expected this spring. More at MacRumors.
Context that matters:
The music industry is already lawyering up. In 2024, major labels sued Suno and Udio over training on copyrighted recordings. Since then, Warner settled with Suno and struck a licensing deal; Universal settled with Udio and announced a partnership. Expect more cases to shift from courtroom fights to licensing contracts. See the RIAA statement.
Market reaction:
Markets noticed. Spotify briefly erased gains and Sirius XM dipped after Google’s launch. Analysts at Bloomberg Intelligence say it’s not fatal for Spotify, but it does raise pressure to add AI mixing and similar features. Coverage at latimes.com.
Why founders should care - the operator take:
Product velocity: Need temporary music for an ad or an app flow? Draft it in minutes, then replace or license later. Lyria 3 is good enough for quick ideation. See the Google blog.
Cost and quotas: Plan around daily caps - free tiers are for demos; paid tiers are for production. Details at latimes.com.
IP risk: Filters and checking tools help, but they are not perfect. If your product touches music catalogs, pursue licenses or label partnerships early. Recent settlements show that deals are possible. More from the Google blog.
Competition and opportunity: Distribution often beats bells and whistles. Analysts think moves like this could push Spotify toward AI mixing. Niche winners may be in stems, mastering, rights management, and creator tools for musicians.
Quick definitions so no one’s lost:
Lyria 3 (Google DeepMind) - Google’s latest generative music model. deepmind.google
Nano Banana - Google’s image-generation model used to create cover art. ai.google.dev
Apple Intelligence - Apple’s system-wide AI layer across iOS and macOS, powering features like Playlist Playground. apple.com newsroom
Playlist Playground - New Apple Music feature that builds a 25-song playlist from a text prompt. MacRumors
Bottom line:
AI won’t replace musicians, but it just ate the music demo phase. Use it to move faster, but get legal and rights advice early.
Get daily insider tech news delivered to your inbox every weekday morning.
