Suno is doubling its valuation to $5B, Spotify opened its platform to AI-generated personal audio, a new study shows consumers prefer AI music until they know it's AI, and Sony's fair-use cases against Suno and Udio are heading toward a pivotal summer ruling.
Suno chasing $5B valuation in Series D, just months after its $2.45B raise
Suno is in talks to close a Series D round north of $250 million that would value the company at over $5 billion — more than double its November 2025 valuation of $2.45 billion. The numbers behind the raise are striking: 2 million paid subscribers, $300 million in annual recurring revenue, and over 100 million total users. Multiple music industry investors are participating, though most are staying anonymous given Suno's ongoing legal battles with Sony Music and Universal Music Group.
For context, Suno's last $250M Series C closed just six months ago. That kind of fundraising velocity says VCs believe AI music generation is now a consumer market, not an experiment. Whether that conviction survives the Sony fair-use ruling this summer is another question entirely.
Spotify opens its platform to AI-generated personal audio
Spotify launched a "Save to Spotify" beta on May 7 that lets AI agents — including OpenAI's Codex and Anthropic's Claude — generate personal podcasts and save them directly to a user's Spotify library. Users write a prompt like "build me a deep dive on World Cup history," an agent generates the audio, and it shows up in your library alongside your playlists. The feature works for both Free and Premium users.
This is Spotify planting a flag: it wants to be the distribution layer for AI-generated audio, not just human-made content. If you build audio tools or create content programmatically, the new CLI tool and its API surface are worth watching. The bet is that personal, on-demand audio is a format category of its own — and Spotify wants to own it before anyone else does.
New research: people prefer AI music — until you tell them it's AI
A study published May 4 by researchers Jana Friedrichsen, Julia Schwarz, and Michel Clement found that listeners enjoy and even prefer AI-generated music in blind tests — but their preference flips once they learn the music was AI-made. More importantly, willingness to pay drops significantly with disclosure. The researchers warn that if streaming platforms become saturated with labeled AI-generated tracks, the total royalty pool could shrink.
If you're a working musician, this is both reassuring and unsettling. The quality gap is closing (or closed), but the "AI" label still carries a stigma that suppresses perceived value. The practical implication: transparency requirements — which are coming in both the EU and US — will reshape how AI music competes for listener attention and revenue.
OpenAI ships GPT-Realtime-2 with production-grade voice intelligence
OpenAI released three new audio models on May 7: GPT-Realtime-2 (live conversational voice with GPT-5-class reasoning and 128K context), GPT-Realtime-Translate (real-time translation across 70+ languages), and GPT-Realtime-Whisper (live speech-to-text). GPT-Realtime-2 scores 15 points higher on audio intelligence benchmarks than its predecessor and supports multi-tool calling mid-conversation. Pricing starts at $32 per million audio input tokens.
This isn't a music generation model, but it matters for audio creators building interactive voice experiences, AI-driven audio tools, or conversational sound design workflows. The 128K context window and real-time translation open up practical applications in multilingual audio production and voice-driven creative tools that weren't feasible at this quality level before.
Sony's fair-use cases against Suno and Udio set for pivotal summer hearing
While Warner and UMG have settled and signed licensing deals with Suno and Udio respectively, Sony Music remains the holdout — and its cases could define the legal framework for every AI music company. Suno filed for summary judgment in March arguing that training on copyrighted recordings is transformative fair use. A hearing is scheduled for July 2026 in Massachusetts federal court. The SDNY case against Udio runs in parallel.
This is the case that matters. If the court rules training is infringement, every unlicensed AI music model becomes legally exposed overnight. If it rules training is fair use, the leverage shifts permanently away from rights holders. Whichever way it goes, this ruling will be cited in every AI-and-copyright case for a decade. If you build with or depend on AI music tools, plan for both outcomes.
ElevenLabs entered the AI music generation race with ElevenMusic, an iOS app that lets you generate songs from text prompts — specifying mood, length, lyrical content, and writing style. You get up to seven songs per day on the free tier, with a Pro subscription ($9.99/month) unlocking 500 tracks per month. The app includes community remixing, mood-based playlists (Focus, Chill, Cosmic, Late Night), and chart-style discovery for AI-generated tracks.
It's a direct competitor to Suno and Udio, but it leverages ElevenLabs' established strengths in voice and audio synthesis. The generation quality is solid, though it's still trailing Suno v5.5 on vocal expressiveness. The real play here may be the integration story — if ElevenLabs connects its music generation to its voice cloning and sound effects APIs, that's a full-stack AI audio platform no one else has.
A new wave of tools is making it possible to create custom VST/AU plugins from text prompts with no DSP or coding experience. Platforms like Amorph from Artists in DSP let you describe a plugin in plain English, generate the code via an AI agent, then compile it into a working plugin you can load in your DAW. The company's CEO openly admits they pivoted from traditional plugin development after seeing how capable AI coding agents had become.
This is still early and rough — you're not going to vibe-code the next FabFilter Pro-Q. But for one-off utility processors, custom sound manglers, or rapid prototyping of audio ideas, the friction has dropped to nearly zero. If this matures, it changes the economics of plugin development and puts bespoke audio tools within reach of every producer who can describe what they want.