Fish Audio Nears 50 Percent Enterprise Mix

Hitting nearly 50 percent enterprise mix by the time Fish Audio reached a $10M run rate shows the company was not just selling creator voice tools, it was already becoming infrastructure inside other AI products. That matters because enterprise voice deals are stickier than subscriptions. They usually bundle higher concurrency, custom billing, privacy controls, and deployment support, which turns a voice API into a production dependency for customers shipping calls, avatars, or real time assistants at scale.1 2 3 4 5

• Fish Audio sells in three layers. Creators pay for subscriptions and credit bundles, developers pay usage based API fees, and enterprises get custom volume pricing and custom rate limits. The docs also show enterprise specific billing and concurrency tiers, which fits the jump in enterprise mix as larger workloads moved into production.1 2 3
• The named partners show where that enterprise demand comes from. HeyGen uses Fish as a selectable speech engine inside avatar video creation. Telnyx added Fish as a TTS provider for Voice AI call flows. These are product embedded channels where Fish gets paid when another platform’s customers generate speech.4 5
• This mirrors the broader voice stack. Orchestration and application companies like Retell, LiveKit, and communications platforms choose speech vendors based on latency, quality, and price. That makes distribution powerful, but it also means Fish has to win recurring traffic inside partner workflows, not just one off signups from individual users.6 7 8

The next phase is a climb from voice model vendor to deeper voice infrastructure layer. If Fish keeps converting partners into enterprise contracts with custom deployment, privacy, and operational controls, more of revenue should come from high volume embedded usage inside video, telephony, and voice agent products, which would make the business less consumer like and more like core application infrastructure.3 5 7 8