Fish Audio, a Palo Alto–based AI voice generation startup, has raised $52 million in a seed round led by Coreline Ventures and Capital Today, with participation from 359 Capital, Parable, Play Time, Alphalist Partners, Bayhouse Ventures, Carya Venture Partners and HF0. The company says it now has more than 8 million users across its open-source and hosted AI models and generates $21 million in annual recurring revenue.
Fish Audio funding and growth
The startup was founded by former Nvidia researcher Shijia Liao and began as an open-source voice-generation project. Since then Fish Audio has released five AI models, four focused on speech generation and one for speech-to-text. CEO and co-founder Rissa Cao said enterprise clients have different priorities: some firms, like HeyGen, value highly realistic voices for AI avatars, while voice agent platforms such as LiveKit seek natural, low-latency speech.
Models, data sourcing and controversy
Fish Audio builds part of its voice library by compensating users who submit their own voices, but the company faced criticism after some creators alleged their voices had been uploaded without consent. Cao said the company has automated its takedown process, reducing removal time to under three minutes. Osuke Honda of Coreline Ventures said durable trust requires built-in consent and attribution systems, and Rico Mallozzi of 359 Capital said Fish Audio’s fine-grained controls and efficient model training position it well against larger, better-funded AI labs.
Featured image credit: Fish Audio
