Fish Audio Raises $52 Million to Innovate and Navigate Challenges in AI-Generated Voice Models

Palo Alto-based Fish Audio is rapidly advancing the market for AI-generated voice models with diverse applications from creative media to enterprise customer support, aiming to make AI voices more expressive and steerable.

• Since its launch, Fish Audio has attracted over 8 million users and generated $21 million in annual recurring revenue.
• On Tuesday, it announced the completion of a $52 million seed funding round led by Coreline Ventures and Capital Today, with participation from several other investors including 359 Capital and Parable.
• Originally started by former Nvidia researcher Shijia Liao, the startup began as an open-source project to address the lack of expressive synthetic voices.
• Fish Audio's portfolio includes five models—four speech generation and one speech-to-text—three of which are open-source, with their latest model requiring a paid API.
• While Fish Audio has made strides, they've faced challenges regarding unauthorized voice use, leading to claims and the implementation of an automated DMCA takedown process.
• Investors like Osuke Honda stress the importance of trust, transparency, and consent within community-driven platforms for sustained success.
• The company plans to launch an audio understanding model and a speech-to-speech model in the future, amid a competitive landscape with peers like ElevenLabs and WellSaid.
• Despite these challenges, 359 Capital highlights that Fish Audio's technical skills position it competitively against larger AI firms.