Fish Audio makes a splash after raising $52M seed funding for AI voices
SiliconANGLE Mike Wheatley ● Covered by 2 sources
Fish Audio just landed $52M for its AI voice tech, going from one guy's bedroom laptop to an 8-million-user platform.
Based on reporting by SiliconANGLE, Mike Wheatley — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Every voice AI origin story seems to start with someone annoyed at how robotic these things sound. Shijia Liao's version began in his bedroom, with a single GPU and a serious dislike for the flat, lifeless text-to-speech models he kept running into. Rather than wait for someone else to fix it, the former Nvidia video researcher and self-described anime obsessive started training his own models. That side project became Fish Speech, an open-source release that racked up more than 31,000 GitHub stars and pulled in indie developers and game designers hungry for voices that actually had some life in them.
Three years later, that bedroom experiment is now Hanabi AI Inc., trading under the Fish Audio name, and it just closed a $52 million seed round led by Coreline Ventures and Capital Today, with 645 Ventures, HF0 and a handful of others joining in. The platform now supports 83 languages, clones a voice from a five-second clip in under 15 seconds, and gives developers control over tone and pacing through more than 15,000 emotion prompts. In blind listening tests, 67% of participants picked Fish Audio's output over rival systems, a number the company is clearly happy to repeat.
The growth numbers back up the hype a bit. Eight million users and $21 million in annual recurring revenue is a serious jump for something that started as a hobby project, and it's enough to put Fish Audio on the radar of companies like ElevenLabs, which just raised $500 million at an $11 billion valuation. Fish Audio isn't playing in the same weight class yet, but it's positioning itself as the scrappier, cheaper alternative — and cheap is a real selling point when the plan is to give away the flagship S2.1 Pro model through its API starting at the end of August.
What's notable is where Fish Audio wants to go next. CEO Rissa Cao and Liao aren't just chasing content creators and game studios anymore; they're courting healthcare and financial firms with HIPAA-compliant, on-premises deployments and zero data retention. That's a very different customer than the one who found this thing on GitHub in 2023. The next moves — voice-native LLMs, real-time speech-to-speech translation, a proper enterprise sales team — suggest a startup trying to become infrastructure rather than just a clever tool, betting that voice ends up being how people talk to every AI system, not just the flashy ones.
My take — AI-written commentary, not fact-checked reporting
Giving away a flagship model for free is a classic move to buy market share before the giants notice, and it usually works right up until the free tier quietly disappears. Fish Audio's real advantage isn't the tech specs, it's the origin story — investors love funding the scrappy underdog who beat well-funded labs with one GPU, and that narrative is worth more to a seed round than any benchmark. Enterprise healthcare and finance clients, though, tend to care less about anime-fueled passion projects and more about who's still around in three years.
Read more about this at: SiliconANGLE