Fish Audio raises $50 million seed round for AI voice models

In short: Fish Audio says it raised $50 million to expand its AI voice models for creators and business customers.

What happened

Fish Audio, a startup based in Palo Alto, said it has raised a $50 million seed round led by Coreline Ventures and Capital Today. Other investors included 359 Capital, Parable, Play Time, Alphalist Partners, Bayhouse Ventures, Carya Venture Partners, and HF0.

The company makes AI voice models, tools that can generate speech that sounds like a person talking (like a very advanced text-to-speech voice). Fish Audio says it has more than 8 million users across open-source and hosted versions of its models, and it generates $21 million in annual recurring revenue, meaning steady subscription-style revenue that repeats each year.

Fish Audio began as an open-source project from Shijia Liao, a former NVIDIA researcher. The company says its “Fish Speech” project on GitHub has more than 31,000 stars, and it is used by developers, game designers, and creators.

Fish Audio has released five models in the past year, including four for generating speech and one for turning speech into text. It open-sourced three speech models, but its latest “S2.1 Pro” model is only available through a paid API (an API is like a menu that lets other apps order the voice feature).

The company also described changes to its takedown process after some creators alleged their voices were uploaded without consent. CEO Rissa Cao said takedowns are now automated, and can remove a voice in under three minutes once a creator proves ownership.

Why it matters

AI voices are showing up in videos, games, and customer service calls. How companies handle consent and fast removal is important, because a voice can be copied and used before the real person even knows it happened.

Source: TechCrunch AI