Fish Audio secures $50 million to expand controllable voice AI for creators and enterprises
Palo Alto-based Fish Audio announced on Tuesday that it raised $50 million in a seed round led by Coreline Ventures and Capital Today. Launched last year, the AI voice startup now has more than 8 million users across its open-source and hosted models and reports $21 million in annual recurring revenue.
Why controllable voice matters
Creative and enterprise applications demand different qualities from synthetic speech. Game studios need expressive characters, AI avatar platforms prioritize realism, and voice-agent providers require natural, low-latency output suitable for live calls. Fish Audio addresses these requirements with more than 15,000 natural-language controls.
The technology also raises questions about consent and ownership. Some creators alleged that their voices had been uploaded without permission. Fish Audio has since automated its DMCA process: CEO and co-founder Rissa Cao says creators can submit a voice sample or contract and secure removal in under three minutes. However, unauthorized uploads may remain available until the affected person discovers and reports them.
Models, customers and development plans
Fish Audio began as a project by former NVIDIA researcher Shijia Liao, who trained a voice-generation model on one GPU and released it as open source. The Fish Speech repository has since collected more than 31,000 GitHub stars. Over the past year, the company launched four speech-generation models and one speech-to-text model.
Three generation models are open source, while S2.1 Pro is restricted to the paid API. Fish Audio sells monthly plans with generation minutes and voice cloning, alongside enterprise APIs used by HeyGen, Sanas and Plaud. The company plans to release an audio-understanding model this year and is also developing speech-to-speech technology.
“A community-centric approach can only become a durable advantage if creators trust the platform.”
Coreline Ventures partner Oskue Honda argues that consent, attribution, licensing and simple reporting must be built into voice platforms. Fish Audio will compete with ElevenLabs, WellSaid, Cartesia, Speechify, Async and Krisp as these standards evolve.
For businesses, the funding signals that voice AI is moving from basic text-to-speech toward configurable production infrastructure. Buyers should evaluate latency, expressive control and cost, but also demand verified voice ownership, clear commercial rights and reliable takedown procedures before deployment.