Google Acquires ProducerAI as Lyria 3 Music Rolls Out in Gemini

The Story

Google announced on February 24 that it has acquired ProducerAI, the AI music platform previously known as Riffusion, and is bringing its team into Google Labs and Google DeepMind. ProducerAI runs on a preview version of Lyria 3 and integrates with Gemini, Veo, and other Google AI systems, with outputs watermarked via SynthID. Google is also rolling out Lyria 3 music creation inside the Gemini app with 30-second generation as a baseline and an 18+ age gate, as scrutiny continues around AI music training data and copyright.

The Facts

• ProducerAI joins Google Labs and DeepMind

  The acquisition was confirmed in a Google blog post by Elias Roman, Senior Director of Product Management at Google Labs. LinkedIn posts from ProducerAI executives indicate the full team is joining Google across Labs and DeepMind.

• Founded as Riffusion, viral in 2022

  ProducerAI was founded by Seth Forsgren and Hayk Martiros. It began as Riffusion, an open-source hobby project that went viral in December 2022.

• Raised $4 million in seed funding

  In October 2023, the company raised a $4 million seed round led by Greycroft, with participation from South Park Commons and Sky9. The Chainsmokers joined as advisors.

• Launched as ProducerAI in July 2025

  The platform relaunched in July 2025 as ProducerAI, initially using its own AI model before integrating with Google systems.

• Powered by Google AI stack

  Under Google, ProducerAI uses:

  • A preview version of Lyria 3, developed by Google DeepMind, which can turn text and image inputs into audio outputs

  • Gemini for its conversational interface inside Google’s flagship chatbot app

  • Nano Banana for album art creation

  • Veo for AI-generated music videos

    All outputs include Google’s SynthID watermark to identify AI-generated content.

• Competitive market context

  ProducerAI enters a market dominated by better-funded competitors such as Suno, which raised $250 million in a Series C round at a $2.45 billion valuation and reports $200 million in annual revenue.

• Legal pressure on competitors and industry lawsuits

  Suno settled a copyright lawsuit with Warner Music Group and formed a licensing partnership, but continues to face infringement suits from Sony Music, Universal Music Group, and European rights organizations including Koda and GEMA. Separately, music publishers sued Anthropic for $3 billion in January 2026, alleging unlawful use of more than 20,000 copyrighted songs for AI training. A federal judge previously ruled that training on copyrighted data may be legal, but pirating it is not.

• Ongoing AI dealmaking

  The acquisition follows other AI-focused moves by Google, including bringing in leadership and engineers from voice AI startup Hume AI via a licensing deal.

Lyria 3 in Gemini (What It Does and How It Works)

• Public rollout and baseline limits

  Google is making Lyria 3 available in the Gemini app in beta, with access rolling out over days. Free subscribers are capped at 30-second outputs, while Google AI Plus, Pro, and Ultra subscribers get higher limits (exact limits not specified).

• Eligibility and language availability

  Lyria 3 music generation in Gemini is limited to users 18+ and is available globally in English, Japanese, Korean, Hindi, Spanish, Portuguese, German, and French, with more languages planned.

• How you use it (Gemini and Workspace)

  In the Gemini app, users select “Create music” from the tools menu. For Google Workspace environments, admins control access via existing Generative AI settings in the Workspace Admin console.

• Inputs and outputs

  Users can generate music using text prompts or uploaded images; Google also describes uploading videos for visual prompting. Gemini returns a track plus AI-generated album art created by Google’s Nano Banana model.

• Quality and control claims

  Google describes Lyria 3 as its “most advanced” music-generation model, aiming for high-fidelity music with natural flow “from note to note,” and says it was developed with input from producers and musicians to better understand musicality (rhythm through arrangement). Google also claims Lyria 3 generates its own lyrics, offers more control over style, tempo, and vocals, and produces more realistic and musically complex tracks.

• Watermarking and detection

  Google says tracks are embedded with SynthID and that users can upload an audio clip to Gemini to check whether SynthID is present.

• Artist usage and Music AI Sandbox

  Google said three-time Grammy-winning artist Wyclef Jean used Lyria 3 and Google’s Music AI Sandbox on his song “Back From Abu Dhabi,” including using Google’s tools to add a flute sound to an already recorded track.

• Training data transparency and skepticism

  Google has not disclosed specific training data sources for Lyria 3. A Google representative said the company was mindful of copyright and partner agreements and that Lyria 3 was trained only on music that YouTube and Google have a right to use under terms of service, partner agreements, and applicable law; public criticism on X questioned whether this explanation is sufficiently specific.

Why This Matters

Google is pairing an acquired, consumer-facing music product (ProducerAI) with a broader public rollout of Lyria 3 inside Gemini, including distribution controls (age gating, Workspace admin settings) and provenance tooling (SynthID plus detection). That combination could accelerate mainstream adoption of AI music tools while intensifying pressure on unresolved questions around training data rights and licensing norms.

This article was drafted with the assistance of generative AI. All facts and details were reviewed and confirmed by an editor prior to publication.