Inside Fish Audio's Breakout Year: How Rissa Cao Built a Voice AI Challenger to ElevenLabs

Rissa Cao didn't set out to raise a $52 million seed round in her first year as CEO. She set out to make voice AI sound human, and built one of the fastest-growing companies in the space along the way, the $52 million seed round raise came as a result of that.

From Open Source Project to Category Contender

Fish Audio is a voice AI company building expressive, human-sounding text-to-speech and voice cloning technology for creators, developers, and enterprises. What began as an open source passion project has, in just one year, grown into a serious challenger in a market long dominated by a handful of players.

Two years before this conversation, Rissa's co-founder and chief scientist, Shijia Liao, was running an open source initiative called Open Audio alongside two other prominent open source contributors. Their release, Fish Speech, became one of the most starred repositories on GitHub, eventually surpassing 30,000 stars. That groundswell of developer interest became the foundation for what Fish Audio is today.

The company itself, formally named 39 AI, Inc, was incorporated on July 28, 2025, and most recently had its one year anniversary. In that single year, Fish Audio moved from serving individual developers to onboarding enterprise customers, a shift that began roughly eight months into the company's life.

The founding insight was simple: even the most advanced AI systems tend to fall flat the moment they open their mouth.

"You can build the smartest system in the world but it could still sound like a machine just reading off a script."

Rissa, Shijia, and Jiahua Liu became convinced that the fix was not a better filter layered on top of existing technology, but a fundamentally more expressive model built from the ground up. That conviction shaped the company's mission from day one.

"Voice is becoming the default way we interact with AI, and we definitely want to build the technology that powers it."

An Unlikely Co-Founder Story

Rissa's path to Fish Audio began with a recruiting pitch that flipped on her. Before starting the company, she was working on growth and go-to-market at a different startup, building on a background that included product roles at Meta Reality Labs and Amazon Alexa, as well as prior co-founder experience at another venture.

Shijia was exactly the kind of research talent she wanted to hire. Instead, he reverse recruited her, convincing her to build Fish Audio alongside him. She describes him as one of the sharpest researchers she has ever worked with, someone who can take an idea and turn it into reality almost immediately.

That speed defines how the founding team operates. Rissa said she and Shijia rarely pause to weigh risk or consequence before acting. They move, they observe the result, and they adjust.

"We just immediately put it in action and do it. If it works well, we are winning. It doesn't work, it's our learning experience."

That instinct extended into her own transition from working inside large organizations to running a startup. At companies like Meta and Amazon, she said, enormous energy goes into internal reviews and layers of leadership alignment. In the earliest stages of a startup, however, the data needed to make a fully informed decision often does not exist yet. Founders, she argued, need to form a gut-level opinion quickly, rally their team, and start executing rather than waiting for certainty that may never arrive.

That approach paid off when skeptics pointed to ElevenLabs, a competitor that had already raised hundreds of millions of dollars at a multibillion dollar valuation, and suggested Fish Audio was too late to the market. By the time the doubts reached her, she said, the team had already tested the idea and proven it worked.

Wearing Every Hat in the Company's First Year

Rissa's role has shifted constantly since Fish Audio's founding. Early on, she focused on growth and go-to-market strategy. As investor interest grew, she found herself learning fundraising and due diligence on the fly.

She took on partnerships, compute procurement, legal and finance coordination, and even financial forecasting using Claude, despite having no finance background. Fish Audio did not have a dedicated finance function until recently. She also entered enterprise sales for the first time in her career, driven entirely by inbound interest from customers rather than active prospecting.

"It was just all different types of things coming together, but it was fun."

The $52 Million Seed Round

Fish Audio's seed round was, by Rissa's own account, larger than she ever anticipated. The company was not originally planning to raise significant capital, believing it could scale profitably on its own. Instead, investor interest snowballed, and she found herself coordinating due diligence and allocation across a growing list of backers, all while operating a company that was less than a year old and lacked the polished growth metrics investors typically expect.

Pulling numbers like MRR, NRR, and GRR out of MongoDB and formatting them for investor scrutiny was, in her words, one of the most challenging parts of the process. Still, backers bet on the team's trajectory rather than a perfectly built financial model.

Investors repeatedly asked how Fish Audio positioned itself against ElevenLabs, described by Rissa as "the elephant in the room," along with questions about profitability plans, capital allocation, and hiring priorities. What ultimately won them over was execution speed. Fish Audio kept shipping, kept growing across creators, developers, and enterprises simultaneously, and kept hitting milestones ahead of expectations.

"Nobody asked me this question anymore as it was maybe like eight months ago. They were still asking, because every single thing we tried, we grow traffic, we grow revenue, we grow users. We have users liking us from every single segment."

The new capital is earmarked for three priorities: continued investment in research and development to keep pace with a fast-moving model landscape, expanded brand campaigns following what Rissa called the largest marketing push of her career, and deeper investment in developer relations and tooling, the community Fish Audio grew from originally.

In the 48 to 72 hours following the round's announcement, the company saw a substantial new pipeline of enterprise interest, and it is now hiring aggressively across go-to-market and enterprise sales roles. Looking ahead, Rissa said the company is targeting $100 million in annual recurring revenue and expects to approach profitability by the end of the year. A future funding round, she said, would likely coincide with hitting that milestone or pursuing a specific model breakthrough that requires additional capital, since research and development remains the company's largest cost center.

Standing Apart from ElevenLabs

Fish Audio's product suite closely mirrors ElevenLabs, serving creators, developers, and enterprises under one roof. But Rissa pointed to several areas where she believes Fish Audio pulls ahead.

The company's models retain accents more faithfully during voice cloning, whether Mandarin-influenced English, Malaysian English, Singaporean English, Hindi-accented English, or Australian and British English. For users seeking neutral, unaccented American English, she conceded ElevenLabs may still have an edge due to a larger dataset. Fish Audio also performs strongly across multiple languages, including Mandarin, Japanese, Korean, Spanish, Arabic, German, French, and Italian.

Rissa credits much of the company's naturalness and emotional expressiveness to real user feedback collected from its creator community, which the team uses to fine-tune emotional nuance and controllability. She also pointed to latency as a differentiator, saying the company has found a strong balance between model accuracy and response speed.

A Market Bigger Than People Realize

Rissa believes most people underestimate how large the voice AI opportunity actually is. She estimates that within the next year or two, audio-based token consumption could account for roughly a third of all AI token usage.

She points to a steady stream of companies switching from providers like Amazon Polly, Microsoft Azure, or ElevenLabs to Fish Audio, alongside an equally significant group of companies adopting voice AI for the very first time.

To support that growth, Fish Audio recently stood up a dedicated enterprise team, bringing together forward deployed engineers, account executives, business development representatives, and research staff who collaborate daily to solve customer problems as they arise. Before this team existed, enterprise support consumed only a fraction of the company's attention. As enterprise contracts grew larger and more frequent, the team became a necessity.

The company also prides itself on shipping speed. Most feature requests from users are fulfilled within days, sometimes hours, unless they fall entirely outside the current roadmap.

Consent and Trust in the Age of Voice Cloning

Voice cloning technology raises immediate questions about consent, identity, and misuse, and Rissa acknowledged that Fish Audio has learned some of these lessons the hard way.

The company built what she describes as the largest user generated voice library in the world, with more than two million community voices. Despite consent and release protocols, some users found their voices cloned and used without authorization. In response, Fish Audio implemented a Digital Millennium Copyright Act takedown process that once took days to process manually and now runs in about three minutes.

The company also launched professional voice cloning last month, allowing verified voice actors to publish their voices publicly and earn a fifty-fifty revenue split whenever their voice is used. To publish a voice, users must either match a verified reference recording or submit ownership documentation.

The motivations for publishing a voice publicly vary widely. Some creators, like a Filipino voice actor Rissa mentioned, want to raise awareness of underrepresented accents and offered his voice for free just to broaden representation on the platform. Others are professional voice actors looking to generate passive income without adding more studio hours. Still others choose to remain anonymous, comfortable earning revenue from their voice without ever attaching their name to it.

Rissa acknowledged this technology is already reshaping how professional content is made. She cited a well known content creator whose editor used a voice clone to correct and republish a video, avoiding a full re-recording.

"The voice overs is much more efficient for content creators to produce content."

She described it as one piece of a much larger shift already underway in professional media production.

Culture, Hiring, and What She Looks For

Fish Audio's most effective growth channel has been an unexpected one: LinkedIn. Since March, Rissa's posts have attracted attention not just from investors, but from go-to-market talent and even employees at competing companies, some of whom reached out directly and later joined conversations about working together.

When hiring, she prioritizes candidates with a demonstrated track record at a similar stage of growth or a clear domain expertise. Beyond experience, she looks for passion and what she calls high agency, a willingness to act quickly without overthinking. She is wary of candidates who lead with requests for advisory equity or executive titles before demonstrating any value. The people she trusts most, she said, are the ones who offer help first and worry about compensation later.

"People who don't overthink are very important if you want to be successful at Fish Audio, or in general startups."

As a leader, she avoids simply telling people they are wrong. Instead, she explains her reasoning and gives actionable feedback, a habit she credits with building trust across a team she describes as highly diversified. She does not believe leadership requires always being the one out front. If another team member has a stronger go-to-market strategy, she said, the team follows that strategy, regardless of who proposed it.

Building a Global Team from Tokyo to Palo Alto

Fish Audio operates with two hubs: a research and engineering team based in Tokyo and a smaller research, growth, and go-to-market presence in Palo Alto. Rissa's prior experience operating teams in Tokyo, combined with a shared team culture rooted in anime and VTubing fandom, made the city a natural fit.

She also pointed to talent economics. Assembling a top tier research team in Tokyo is more affordable than doing so in Silicon Valley, where researchers frequently move between labs like Gemini, OpenAI, and Thinking Machines chasing compensation. In Tokyo, she said, the team is bound more tightly by shared mission than by negotiation.

Navigating Entrepreneurship as a Female Founder

Asked directly about navigating a male dominated industry, Rissa did not shy away from the topic. She has encountered people, particularly in male-dominated founding teams, who try to undermine her confidence. Her response has been consistent throughout her career: she does not internalize that doubt.

She traces some of that resilience back to her upbringing, describing parents who never punished mistakes and instead treated failure as an opportunity to learn. That mindset has carried directly into how she leads.

Advice for the Next Generation of Founders

For students eyeing the startup world, Rissa's advice is consistent with how she runs Fish Audio itself: stop overthinking and start doing. She treats her own interns the way she treats herself, giving them a high level playbook and then letting them operate, fail, and adjust in real time. Perfection on the first attempt, she said, is not realistic for anyone.

"Unless you do it, you are iterating on the last version of yourself. Otherwise you can't do something perfect at the first try. Nobody is able to do that."

Looking Ahead

A year into Fish Audio's life, Rissa is candid that the company's long term definition of success will keep evolving alongside a fast-moving industry. What has not changed is the team's appetite for breaking its own records, week over week, whether in model performance, revenue growth, or the breadth of creators, developers, and enterprises it serves.

With fresh capital, a growing enterprise motion, and a research team she considers among the best in the world, Fish Audio heads into its second year with the same instinct that built its first: keep shipping, keep listening to its community, and let the results speak for themselves.