In short
The episode “Monetizing AI: Fish Audio’s $52M Strategy” discusses how AI voice model startups make money, using Fish Audio (Phish Audio in transcript) as the main case study. Topic: Fish Audio raised $52M and reportedly has ~$21M annual recurring revenue; it has 8M+ users across open-source and hosted models, and uses open-source adoption to drive enterprise licensing.
Guest backgrounds
The transcript names John Daly (host) and Jaden (co-host) plus Fish Audio CEO/co-founder Risa Cao and founder Shijiao Lau (former NVIDIA researcher).
Key claims
open-source “free” models build traction; GitHub stars (31,000) can be gamed/overvalued; enterprises pay for licenses while indie developers can demo for free.
Notable examples
HeyGen (realistic avatar voices) vs gaming studios needing expressive character voices; the hosts also claim they licensed their podcast audio for AI training and earned thousands of dollars, and cite the Claude $1.5B lawsuit as a compensation opportunity for content used in training.
Written by AI. May contain mistakes. Listen to the episode to check what was said.
Chapters
Tap a time to open that second in VOPhish Audio's $52 Million Strategy
1:00 to 2:38
Discussion about Phish Audio's funding and business model.
“Today we are talking again about AI voice models.”
Open Source vs Closed Source in AI
2:40 to 4:22
Exploration of Phish Audio's strategy with open source models.
“with this story that you're mentioning with Phish Audio because$52 million is a ton of money, but they did have a lot of annual recurring revenue.”
Opportunities in AI Voice Licensing
4:25 to 7:53
How to monetize audio content and the potential for creators.
“I think this is an interesting company because in my mind, there's so many voice models out there already.”
Transcript
Automatic transcript. May contain errors.0:28John Daly here. You're listening to a podcast right now, driving, working out, walking the dog. If you're into podcasts, chances are you have something to say too. With RSS.com, starting your own podcast is free and easy. Upload an episode and we distribute it to Apple Podcasts, Spotify, Amazon Music, and more. Track your listeners, see where they're from, and start earning from ads just like this. If you've been thinking about starting a podcast, this is your sign. Start your new podcast for free today at RSS.com. Welcome back to the podcast. Today we are talking again about AI voice models. This new one called Phish Audio raised$52 million, giving it a, or at a$21 million annual recurring revenue.
1:13So it's another AI voice model startup. They just raised a ton of money and there's a ton of money being made in a lot of different ways in this area. So we're going to break that down for you today. Talk about how we've even licensed this podcast to train AI models and made money. So we're going to get all into that. But before we do, Jaden, why don't you tell them about our school community? Yeah, every week we record a bonus piece of content over on our school community. This week we did a whole deep dive into my Facebook analytics. I spent$2 ,000 running Facebook ads for AI Box in the last week or two.
1:50And so I break down what the results were, what I learned, how I'm scaling that, the winning ad formats. I mean, I literally show you guys all of the numbers and everything that I don't share post publicly. So if you're interested in seeing that, go join the school community. It's$20 or it's$19 a month. You get access to that plus tons of other stuff, including by the way, I got to mention this. If you have any interest in doing any sort of audio, any sort of audio work, there's a way to get$6 ,000 in free 11 labs credits. So$6 ,000 for free from 11 labs. And I also have a whole video explaining how you can qualify and get that as well.
2:28So if you're interested, it's all in the school community. It's also free to join, but if you want to get all of the bonus content, it's$19 a month. There's 600 members and we'd love for you to be one of them. Okay, let's talk about what's going on with this story that you're mentioning with Phish Audio because$52 million is a ton of money, but they did have a lot of annual recurring revenue. They're based in Palo Alto. They have more than 8 million users across their open source and hosted models. I think that's a key. It's really smart for them to make an open source model, release it. They probably got a ton of usage on that open source model, which they of course use to build their closed source model.
3:06But then when they go and raise money, they're like, hey, look, we have 8 million users on our open source and closed source. So anyways, it's kind of fun. They do the free product that's really popular and they can use that metric. I think that's genius. And if you're raising money, making some sort of really valuable open source tool for people, I think is a really cool way to go. This is their repository. It has 31 ,000 stars since it was launched last year. One interesting caveat that I will put on the GitHub stars, people literally raise money based off of how many stars their GitHub project has.
3:40And I am not advocating fraud. I'm more just like bringing awareness to it. You can go and buy GitHub stars. And sometimes I wonder if some of these companies will go and like put out the project on GitHub, buy 50 ,000 GitHub stars. And then like that kind of like starts the flywheel where like everyone's like, oh wow, a lot of people like are starring this. So anyways, the stars do not matter. The traction, the action, the community built around it, that's what actually matters. Hopefully investors are smart enough to see that. But yeah, whenever I see a call out like 31 ,000 stars, I'm like, I mean, basically means nothing.
4:17But if it's all legit people, then it's a good indication. Right now, their library, I'll go for it. I'll go ahead. I think this is an interesting company because in my mind, there's so many voice models out there already. Why would this one be able to jump in and claim that amount of market share? But I guess this is sort of their angle. The CEO mentioned that there's companies like HeyGen, which they're optimizing the voice AI to power your AI avatar who wants realism and professionalism, but it sounds like their angle is more to make it open source for video game designers, creators, indie developers.
4:58So, and they want to have more expression. That's kind of like what their claim to fame is more expressive voices that maybe a little bit more outrageous that you could apply to a character in a game. So I think that's a really great, a really interesting business model or a way to differentiate because, you know, I think back to, this just came to my mind, but Grand Theft Auto, the video game, they had like these different radio stations you could flip through in the game while you're driving your car and you could, there was like a chat, like a, yeah, chat radio or whatever you call it, just talk radio.
5:38And they would have these ridiculous stories, but they kind of made, they're just kind of really fun, funny to listen to. Um, you could kind of take that if you took the AI voice models, you could almost have it go endlessly, you know, cause after a while on the game, you'd hear repeat stories, but, um, to have like endless talk show radio at while you're driving your car in a video game is, I think a pretty cool use case and it would make the game more interesting. A hundred percent. There's so many different interesting use cases like that. Um, so yeah, really impressive company. It was, uh, it was started, it was basically a side project by former NVIDIA researcher, Shijiao Lau.
6:15And they trained a voice model on a single GPU. And then, you know, based from that, they kind of open source, a lot of people liked it. And then they decided to spin it up into this whole company. And yeah, it's been really popular. Now, you can also imagine why this has been so popular. Risa Cao, who's a CEO and co-founder of Fish Audio said, every enterprise has different use cases and different preferences. For example, companies like HeyGen, which use our voice to power AI avatars, want realism in voices. A gaming studio would want expressive voices for their characters. So there's tons of different use cases.
6:48But I think being able to make something that's like if they have an open source model as part of this, obviously, the enterprises are going to have to pay a license. So they're not just getting it for free. But a lot of indie developers, people can make demos with this for free. I think that's really cool. And when it comes to making money from all of this kind of stuff, the one thing that I will say is if you have any sort of audio library yourself, companies like Phish Audio and lots of others need to go and get high quality data sets to train their model on to make it sound realistic. And so, for example, even this podcast that you're listening to, AI Hustle, we've licensed and made thousands of dollars licensing our audio to be used for not like voice cloning, but general purpose AI training to a lot of these different AI companies.
7:36And if you don't have an exclusive, you can resell it over and over again. And yeah, you can you can make a ton of money. So I think right now is like the is like there's no better time to start creating content, whether that's video, whether that's podcasting. And I think in the future, it's gonna be more like motion capture with video, but there's so many different areas where you can make a lot of money doing audio. My eyes have been open to this recently, mostly because of what we did with our podcast. But because of the recent lawsuit with Claude having the$1.5 billion payout because they pirated every book ever made to train their model, there's a lot of new opportunities to make money coming up.
8:19Because now, if a model is going to train off your content, you deserve to be compensated for it. So again, another reason to join the school community to find out how we've done this and how much money we've made. So anyhow, thanks for listening, guys. If you have any questions or if you are listening to us wherever you are, please leave us a rating or review. We really appreciate those. They help us reach more people. And again, check out the AI Hustle School community so you can learn how to use some of these tools to actually make money online and grow your business. Thanks for listening and we'll see you next time.
8:51moto casino is pure american energy born in the usa and with the same 1776 spirit that built this country moto america's social casino is celebrating that freedom with a 1 million sweeps coin prize giveaway one lucky american is going to win 1 million sweeps coins so what are you waiting for spin now on moto casino download the moto casino app today moto casino is the social casino board prohibited no purchase necessary visit moto.us for more details Motor Casino, America's social casino. You're listening to a podcast right now. Driving, working out, walking the dog. If you're into podcasts, chances are you have something to say too.
9:31With RSS.com, starting your own is free and easy. Upload an episode and we distribute it to Apple Podcasts, Spotify, Amazon Music, and hundreds more. Track your listeners, see where they're from, and start earning from ads like this, even with just 10 listeners a month. If you've been thinking about starting a podcast, this is your sign. Start free at rss.com.
From the publisher
In this episode, we break down how Fish Audio is leveraging their $52 million funding to maximize profits through AI voice technology. Learn innovative ways to utilize AI in your own ventures.




