Google Acquires ProducerAI - The Riffusion Successor That Hit 1 Million Users
Google acquires AI music startup ProducerAI (formerly Riffusion) and folds its team into Google Labs and DeepMind, pairing the platform with Lyria 3 to compete with Suno in the AI music generation market.
7 min read
Google has acquired ProducerAI, the AI music creation platform formerly known as Riffusion, bringing the startup's full team into Google Labs and Google DeepMind. The deal, announced on February 24, pairs ProducerAI's conversational music creation agent with Lyria 3 - Google DeepMind's most advanced music generation model, which launched just six days earlier.
TL;DR
- Google acquires ProducerAI (formerly Riffusion), folding the team into Google Labs and DeepMind. Financial terms were not disclosed
- ProducerAI will be powered by Lyria 3, Google's latest music generation model, enabling tracks up to 3 minutes with vocals, lyrics, and instrumentals
- The platform had 1 million monthly users and $4M in seed funding before the acquisition - now it competes directly with Suno ($2.45B valuation, $200M revenue)
- ProducerAI continues operating at producer.ai with free and paid plans in 250+ countries, but Google retains a perpetual, royalty-free license to all user-generated content
The Deal
The blog post announcing the acquisition was written by Elias Roman, Senior Director of Product Management at Google Labs - a name that should ring bells in the music tech world. Roman co-founded Songza, the music streaming service Google acquired in 2014. He then spent five years managing Google Play Music and YouTube Music before moving to Google Labs. The fact that Google's music acquisition playbook is being run by someone who has done this before is not a coincidence.
What Google gets
ProducerAI is not a prompt box that spits out a song. The platform is built around an iterative, conversational agent that guides users through the music creation process - workshopping lyrics, refining melodies, experimenting with genres. As the Google blog post describes it, ProducerAI is designed for "the back-and-forths that play out over time."
"It's not a tool that you put in your prompt, roll the slot machine, and something will come out."
- Google Labs product blog
Google also gets 1 million monthly users, a conversational music creation UX that took three years to build, and a team with deep technical and musical credentials.
The Riffusion origin story
ProducerAI started life as Riffusion, an open-source experiment released on December 15, 2022, by Seth Forsgren and Hayk Martiros - Princeton classmates who spent over a decade playing in an amateur band together. The original idea was brilliantly hacky: fine-tune Stable Diffusion 1.5 on spectrograms (visual representations of audio), generate spectrogram images from text prompts, then convert them back to audio via inverse Fourier transform. Making music by generating pictures of music.
The demo went viral. A Hacker News post pulled over 2,400 upvotes. Within months, 500,000 tracks had been generated. The founders incorporated as Corpusant, Inc. in San Francisco and raised a $4 million seed round in October 2023, led by Greycroft, with The Chainsmokers coming on as advisors.
The evolution was rapid: lyrics and vocals in late 2023, a mobile app with image-to-song in early 2024, the proprietary FUZZ model in January 2025, and the rebrand to Producer.ai in July 2025 with FUZZ-2.0 supporting multi-minute tracks. By the time Google came calling, the spectrogram hack had become a serious platform.
What Lyria 3 Brings
Lyria 3, launched on February 18, is a fundamentally different beast from Riffusion's original approach. Instead of generating spectrograms - visual proxies for sound - Lyria 3 works directly with audio tokens, compressed units of data that represent music natively. The model generates vocals, lyrics, and instrumentals across multiple genres, with granular controls over tempo, dynamics, and time-aligned lyrics.
The key numbers:
| Feature | In Gemini App | In ProducerAI |
|---|---|---|
| Track length | 30 seconds | Up to 3 minutes |
| Lyrics | Auto-generated | Conversational workshopping |
| Vocals | Multi-language | Multi-language |
| User control | Text prompt | Iterative agent |
| Safety | SynthID watermarked | SynthID watermarked |
The tech stack
Post-acquisition, ProducerAI integrates several Google AI systems:
- Lyria 3 for core music generation
- Gemini for the conversational chat interface
- Veo for AI music video generation
- Nano Banana for album art and cover art
- SynthID for imperceptible audio watermarking on all outputs
The platform also features "Spaces" - a natural language interface for creating custom instruments and effects - and a node-based modular audio patching environment for users who want lower-level control.
The Suno Problem
The elephant in the recording studio is Suno. The AI music startup raised $250 million in November 2025 at a $2.45 billion valuation, reports $200 million in annual revenue, and claims nearly 100 million users. ProducerAI's 1 million monthly users and $4 million in total funding look modest by comparison.
"We're just scratching the surface of what these models are going to be able to do once we harness everything that Google brings to the table."
- Seth Forsgren, ProducerAI co-founder and CEO
Google's resources change the math. Lyria 3 running on Google's infrastructure, distribution through the Gemini app and YouTube's Dream Track, and the DeepMind research pipeline behind it all - these are advantages that no venture-backed startup can match. ProducerAI's conversational agent approach also differentiates it from Suno's more immediate generate-and-download model.
The copyright landscape
The acquisition lands in a minefield. Major labels sued both Suno and Udio in June 2024 for mass copyright infringement. Warner Music settled with both companies by November 2025, and Universal settled with Udio the month before - but Sony's cases remain active. In 2024, hundreds of musicians including Billie Eilish, Katy Perry, and Jon Bon Jovi signed an open letter calling on tech companies not to undermine human creativity with AI music tools.
Google's approach has been notably more cautious. SynthID watermarking is embedded in every generated track and can be verified by uploading to Gemini. The platform includes filters that check outputs against existing content, and Google has collaborated directly with artists - including Grammy-winner Wyclef Jean and rapper Lecrae - on guardrails.
What They Left Out
The fine print deserves scrutiny. Google's terms grant the company a "perpetual, royalty-free license" to all user-generated content for platform operations, promotion, watermarking, and model improvement. Users get commercial distribution rights but not traditional master ownership. For hobbyists making tracks for fun, this is irrelevant. For anyone hoping to build a music career on AI-generated output, the distinction between "commercial rights" and "ownership" matters.
There is also the Google product lifecycle question. The company's history of launching and killing products - from Google Play Music itself (shut down in 2020) to Stadia, Inbox, Google+, and dozens of others - is well documented. ProducerAI's community has noticed. The reaction in ProducerAI's Discord was described as "measured, mixing creator excitement with structural nervousness."
"We are so grateful to see how this platform continues to evolve. It's truly crafted around the musician's experience. The founders are incredibly technical, but natively musicians, and understand the nuances of what makes a platform truly be an additive tool in the creation process."
- Alex Pall, The Chainsmokers
The Chainsmokers' endorsement is nice. But the real test is whether Google treats ProducerAI as a long-term product or as a talent acquisition that gets folded into the Gemini ecosystem and quietly sunsetted once the useful parts have been extracted. Elias Roman's track record - he has been building Google's music AI strategy for over a decade now - suggests the former. But Google's institutional track record suggests hedging your bets.
ProducerAI is available now at producer.ai with free and paid plans (starting at $6/month) in over 250 countries. The Lyria 3 integration is live, with 30-second generation in the Gemini app and tracks up to 3 minutes in the full ProducerAI platform. Whether Google's deep pockets and infrastructure can close the gap with Suno's 100-million-user head start is the question that will define whether this acquisition was a strategic masterstroke or an expensive experiment. The AI image generation market already showed that Google can compete when it commits. Music is the next test.
Sources:
