Transcription
I love living in the present, man. It's great. Like, we're talking about real data centers in space. It's so cool.
I mean, how do we not have a base on the moon 55 years later? It seems like we should. And if you asked anybody who watched the moon landing, hey, when do you think we'll have a base on the moon? They'd be like, I don't know, 10 years, 5 years. I mean, we're going to go up every year now, right? And it's like, no, we're just never going to go back. It has no importance. I I want to know what the hell's going on. I hope I live long enough to understand some about the nature of the universe.
This Week in Startups is brought to you by Kota. KOD empowers your startup by bringing words, tables, and teams together. Strategize, plan, and track goals effectively with all your valuable data in one place. Go to kod.io/twist io/twist to get started for free and get six free months of the team plan.
Oracle. Oracle Cloud Infrastructure or OCI is a single platform for your infrastructure, database, application development, and AI needs. Save up to 50% on your cloud bill at oracle.com/twist.
And Stripe Startups. Stripe Startups offers early stage ventureback startups access to Stripe fee credits, expert insights, and a focused community of builders. Apply today on stripe.com/startups.
All right, everybody. Welcome back to This Week in Startups. It is Monday, October 6th. Oh my god, the summer is over and there is uh snow in the mountains, so I should be able to uh get on my skis next month. Very excited about that. With me again, Alex Wilhelm. How are we doing, Alex?
Fantastic. We were just discussing the importance of deer hunting before the show and I just want to say it on record. Summer sausage is the reason why everyone should go hunt deer because it's the best thing in the world and was a key food ingredient to my youth.
Oh, really? Summer sausage. Okay, good to know. You know, I uh living here in the great state of Texas frequently people I'll see them and they'll hand me a bag. Oh, great to see you. I I brought you some sausage. I I went boar hunting. I went deer hunting. I you know, whatever. And uh people just hand you meat. Uh it's the nature of the place. like people are wearing guns on their hips, they're handing each other meat, I'm chewing on a cigar. I love it. I love it here. Uh really enjoying my second year in the great state. Uh yeah, deep into my second year. And uh let's get started. I saw Open AI and AMD are doing a deal and this deal comes on the heels of just two weeks ago or last week, Jensen putting a hundred billion into
Open AI. So, I'm confused. Tell me everything.
Okay, so we're talking a lot about Project Stargate, Jason, and all the deals that are part of Open AI's, the American AI model company's infrastructure push. This is part of that. You are correct. A couple weeks ago, they announced a 100 billion dollar deal with Nvidia. That appears to be kind of like a share trade for uh hardware. The AMD deal is a little bit different and it could be smaller. So, the gist is this. AMD is going to supply open AAI with up to six gigawatts worth of GPUs, not all at once, in several tranches. The first one will land in the back half of 2026 when their new uh GPU comes out. I think it's the 4150 in their line of GPUs. As part of this deal, OpenAI gets to purchase stock in AMD at 1 cent per share, the par value of its stock in tranches again. So every time it buys a bunch of GPUs, it gets to buy a bunch of their stock for 1 cent per share. About 10% of the company is the way this one's going to shake out. But it's a lot of GPUs. It's a lot of stock. And soon Sam Alman will own a sizable chunk of AMD, the largest of any single shareholder in the world.
Well, OpenAI will own uh not Sam personally, obviously, just to make that super clear for the audience as well.
Fair enough. Yeah. Um and so this is very non-traditional to say the least.
Yes. Yes. So, OpenAI is buying a bunch of compute from AMD. Great. The money they're using to buy that is not AMD's money. AMD is just a supplier to them. AMD is not investing in Open AI.
Unlike Nvidia, which did invest in Open AI.
Correct. That's my understanding of this. We do have an SEC filing in the case of the AMD offering, which is where we got the 1 cent per share point. And as far as we can tell, that's exactly correct, Jason. But no matter how you slice it or how the money moves around, it's OpenAI securing quite a lot more compute capacity for the next couple of years. So it all kind of fits in the same direction. They're just tinkering with the right methods to figure this out. But I just didn't see OpenAI buying a 10% stake in AMD. I just that's that that was not on my bingo card, if you will.
Yeah. So, if they're buying all these chips, it says here in our notes, um, that they could be spending as much as, I don't know, $90 billion in AMD hardware.
So, where does that money come from? Well, Nvidia just said they would invest a hundred billion into Open AI. OpenAI obviously has other investors. So did they take money from Nvidia and Jensen and then invest it into and then are going to buy AMD shares and their shares went up like 40% today right like the market had a huge reaction.
Huge reaction but going back to the Nvidia OpenAI announcement that the text is to support this deployment including data center and power capacity Nvidia intends to invest up to 100 billion in open AAI as the new NVIDIA systems are deployed so essentially they're going to buy stuff from Nvidia and then I think Nvidia gets shares in OpenAI versus in the AMD deal. They get in they're going to buy GPUs from AMD and then they get stock in AMD. So I think it shows the relative power balance between Nvidia on one side lots and AMD on the other side less.
Yeah. Um this is bizarre and um
Tell me why. Tell me why. Well, obviously Nvidia and AMD are in competition and so it's fine if you have a ton of money and you go out and you start buying from two different suppliers. I wouldn't expect anything else, right? You want to have a competitive marketplace. So that makes total sense, right? You don't want all your eggs in one basket if you are open AI. But the thing I find weird is the equity components of these deals. So Nvidia investing a ton of money and buying shares in open AI and that open AI open AI is obviously buying a ton from Nvidia that screams of uh roundt tripping and I'm sure that they are well aware of that and have been very thoughtful about outlining that and then you have this deal which is hey we'll buy some from you and we get some warrants. That's not unheard of but it certainly is non-traditional. It just feels like there's a lot of money circulating around and the most critical interpretation of this is that uh people are investing into open AAI uh and they're buying more revenue and buying revenue and roundtpping is a big no no. So, I think that this is going to create a lot of interesting um you know, knocks on the door from like the SEC and places like that to just understand what's going on. And OpenAI wants to go public next year. So, if they do go public, a lot of this information is going to be shared. But this is all based upon OpenAI successfully becoming a for-profit company.
Sure. And that hasn't happened yet. So, this is like I don't want to say it's a house of cards, but there are a lot of interesting variables here. It feels very non-traditional. It's um yeah, what could go wrong?
The other side of that coin though, just thinking out loud with you as you talk about this, OpenAI is a non-traditional company in its scale and how much it's changed I I would say the world. And it's also facing a what appears to be a chronic compute drought somehow still this far into the AI cycle. So could it solve a non-traditional problem using traditional methods of raising money and so forth? Probably not. But I absolutely agree that what we're seeing here increasingly sounds circular, risky, and also like it's a form of of just leverage, Jason, or risk. It feels like quite a lot could go wrong in many companies if demand for AI services doesn't yield the amount of usage that OpenAI is pre- buying capacity against because the AMD deal doesn't really kick off until the back half of next year when these new GPUs come out that AMD is so proud of.
That is another good point is what if um they don't need this compute? What if uh it levels off you know the impact of compute on the product getting better?
Mhm. That's actually a very interesting question a lot of people are having. We've made this analogy many times before but it feels similar to the dotcom buildout.
Yeah. Of uh fiber and they built up so much fiber and there you know the music industry the movie industry wasn't allowing their content to be streamed. There was no YouTube when that was being built out. So, and storage was wildly expensive. Transport of uh video across that fiber was still pretty expensive. So,
Yeah, this this is probably all um we might be building out far more capacity than we need if all these deals do um you know, come to fruition.
Okay, there's so much exciting innovation happening every day in AI and so many companies are just shipping all these amazing new tools. But hey, these AI startups pose several specific challenges. They require a lot of flexibility and an infrastructure that can scale up fast. Thankfully, Stripe startups is here to help. Yes, our good friends over at Stripe are offering early stage ventureback startups access to a whole suite of perks. Stripe fee credits, insights and mentorship from experts in your field, tools that help you reduce decision fatigue and stay focused, and membership in an active community of real builders. More than three out of four startups in Forbes's AI50 are already using Stripe for payments, billing, invoicing, and taxes, giving you an overall view of your revenue so you can make smarter, more informed decisions. Learn more and apply for the program today by visiting stripe.com/startups. That's stripe.com/startups.
Or uh maybe super intelligence, you know, takes off and we have so many jobs to run to cure cancer, to build bigger bridges to solve math equations that we'll use it and and that's that um Javans paradox, I guess.
Jevans. Yeah. Jeans paradox that people keep talking about uh like the induced traffic paradox. So, it's going to be a while to see. And um I think I heard on CNBC today they said 80% of the gains in the stock market this year came from the AI trade. I don't know if that's exactly true or how they're calculating it,
but my lord um you know there is a lot of um AI spending going around.
Yes. Uh as we record this Jason Open AAI is having its dev day. Thank you for everyone who tuned in to us live in uh in contribution of sticking around with Sam. Uh but they did announce as part of that live stream that chat GPT now has 800 million weekly active users. Jason so plus 100 million from the last milestone at least for now implied there's plenty of capacity uh needed to be built out.
So and uh for those of you wondering producer Claude gives us a great uh and you can pull this up Alex if you have a moment. Uh, Javvon's paradox, J E V O Ns is a counterintuitive economic principle that shows how improving efficiency can actually increase total consumption of a resource rather than decrease it. Named after the 19th century economist William Stanley Jevans, uh, who observed it with coal use in England, the paradox works like this. When you make something more efficient to use, it becomes cheaper to use, the lower cost then triggers several effects. Direct rebound, people use more of it because it's cheaper. Indirect rebound, the money saved gets spent on other things that also consume resources. So your fuel savings might fund an extra vacation flight, which also uses fuel or a second home. Economywise effects, the cheaper resources enables new applications and industries that weren't economically viable before, expanding total demand. And air travel would be an example of that. It just wasn't it wasn't cost-ffective and then it became so cheap that now, you know, Spirit Airlines can take you anywhere. So, there you have it, folks.
Yeah. And if you want to use Claude like we do, head over to claude.ai/twist. You can't do everything that we do without a paid plan. But Twist listeners get 50% off their first three months of Claude Pro. Go to claude.ai/twist to get started.
Um, all right, Jason, let's move on from the open AAI buying the entire world and instead talk about one of your favorite companies, Tesla, which has a new teaser video out that has a very interesting sound effect in the background. I'm going to pull that up now and I want you to tell people what you're hearing in this clip.
I'm listening deeply and uh this is like some sort of a fan blade. It looks like it's spinning. You know those little fans you have in your computer, but it looks bigger and it says 107, which I guess would be tomorrow there'll be an announcement.
I think I know what this is. Not because Elon told me or I got some intel. Uh the Roadster 2 is, I believe, coming this year. uh or in another seven. This is like the vanity project of all vanity projects and it's super late on Elon time. You know, he's frequently late, as I always tell people, frequently late, but never wrong. Uh he always gets it done in the end and and you know, typically does it in in such extraordinary fashion that your mind is blown. So, he did say at some point he was going to create a leaf blower, I think, as a joke, a quiet one, because he was so annoyed with leaf blowers outside of his house that are gaspowered. and we have those in the market. So, I don't think it's a leaf blower. Um, in order to drive really fast, um, you need to have downward draft and there's a number of ways to do that. Typically, the air coming into the front of a car, uh, can go through different vents and push pressure down or you have a spoiler at the back which will push pressure down. And sometimes you'll see as in F1 racing, as uh, we've been talking about on the show, you know, a car goes flying, right? um wind gets underneath it. Not a good situation. There is um an EV hypercall car called the Mick Merry, if I if I'm pronouncing it correctly. M C M U R T R Y.
Hypercars are cars that cost over a million dollars and can go typically well over 200, maybe even 250 miles per hour. Uh and uh you know this is like the higher level than your typical Ferrari or Lamborghini.
Oh yeah. And so this car and we'll show a video of it uh can literally drive upside down. And the reason it can do that, this milliondoll car, is because it has fans underneath it that create that downward draft. So these fans are pulling the car down is the my understanding of it. And here they're showing the car with those fans on and they're slowly rotating it. It says it's got 2,000 kilograms of downward ondemand force uh from zero miles hour. So why is this important? If you're making some insane turn, you're going to grip the road. That's what the tires are for as well, obviously. So I think this could be the downward drift fan for the uh new Roadster 2.0. You know, the whole the whole team was just dying watching this clip right now because if their car drops, they have to build a new one. So, can you imagine this?
Well, there's also a driver in there.
Oh, yeah. But he's fine. He's He's fixable. He's replaceable. This car cost $1.27 million.
No, it's driving forward and backwards. It just You saw it drove a little bit forward.
Mhm. That's insane. It's absolutely
I can't even believe they did this test. But
so prepping for this, I got to kind of poke around the um Mc Merry uh universe, if you will. That car, by the way, Jason's called the Spear Lane. Uh and it's just about a million pounds or about 1.25 million USD. They took it around the Top Gear test track from the the British television show. And not only did it set the fastest time of any car ever, including an old F1 car, it beat that by like five seconds.
Like it was it's crazy fast. Yeah. uh the speed in these cars. There's um I don't know if it was Xiaomi or one of the electric car makers just went over three well over 300 miles per hour in a car as well. So top speeds are going up. I think the top speed records though will ultimately be a combination and the and the winners are going to be hybrids where you have um like the Corvette ZR1 X which I'm getting one of uh when they are not a 100 over sticker. Um, that car has a small, like the Corvette -ray, they took the lesson from the Corvette. It has a small uh electric uh motor and battery. Can only go 10 miles on the battery, but it's just used to get you off the line really fast. Yeah.
And then you you kick into the to the uh ice engine. So,
Jason, I'm going to show you just a really short clip here of the BYD Yang Wang 09 Extreme hitting 308 mph. This is in kilome/h in the top left if you're watching the video. Uh, but what I like about this is just how the world seems so silly at this speed. Like the trees don't look real anymore as it accelerates past 500 km/h.
Don't recommend going that fast.
Yeah, don't don't do it. Especially if you're on a motorcycle near my house at night. Stop that.
Be careful, folks. Be safe. But yeah, so that that's my best guess. The other guesses people had were the leaf blower. Um, it could also be um something to do with HVAC. He's he's been building the HVAC systems in the cars to be super efficient and HVAC systems like air conditioning and heat. These things have not advanced as much as they should. And he was talking about possibly building in one of the calls an HVAC system for like homes or for mobile homes, etc. So maybe there's an electric HVAC system coming.
Don't like the Model Y.
Yeah. Don't some Teslas though have that like um pandemic mode when it like doesn't do outside air transfer. So that falls in the same kind of bucket to me.
Um, yeah. Just efficient HVAC. What they found was the HVAC and ICE cars were never optimized because you just burn more gas. Who cares? Uh, you know, gas industrial complex wants you to burn gas. Why why make it more efficiency when you have a battery and you're trying to get miles and there's mile range anxiety. You look at every part of the car and you optimize and the most inefficient part of the car aside from the drag caused by air is um the HVAC system.
Yeah. And that's why they would say like, "Oh, if you're running out of battery, turn off the HVAC."
Well, they made their own. And yes, one of the features uh the biodefense feature means it pushes uh air through very thick filters that can take out all those particles. It's quite nice actually. I don't want to ruin your day, but have you taken a look at your cloud computing bill lately? They probably give you some kind of deal to start, but over time those bills start to add up. Well, our friends over at Oracle Cloud Infrastructure want to help you cut your cloud bill in half. Yes, that's right, 50%. While you're getting better performance at the same time. OCI is a next generation cloud designed to work with any application, including AI. It's faster, it's more secure, and you can do it for less. We're talking complete cloud infrastructure and services regardless of your specific setup or workload. And OCI costs significantly less than other clouds with a span of 50 interconnected cloud regions and more than 150 OCI services a piece. So you can access your cloud from anywhere and keep your prices consistently low worldwide. So join modal sky danceance animation and more innovative AI tech companies who upgraded to OCI and saved. See if you qualify for half off at oracle.com/twist. That's oracle.com/twist. This offer is only for new US customers with a minimum commitment.
Especially if you have allergies here in Texas, it's kind of a a necessity. but also keeping up with the man, some of these places like Arizona and and Italy and and Texas, it gets so hot when it's 110 degrees. It's hard for your Tesla to keep up. And what they do here is, you know, every all these cars have glass roofs now. Everybody buys $20 inserts on Amazon to block the roof so you just get less sun into your car so your air conditioner can keep up during July and August. The other option is to leave Texas in July and August. That's the option I'm taking most summers to try to get out of here for 6 weeks.
I mean, it's it's kind of like an inverse leaving New York for Florida in the worst of winter. You just got to get out of the heat.
All right. One of my favorite startup categories in the world, Jason, is space as a big science fiction guy. And one of the coolest companies that I've heard of is a company that wants to put data centers not in the ocean, not on land, but up in orbit. And you flagged a great uh clip by Jeff Bezos talking about just that, which I have pulled up. I thought we would listen to it. and hear what my uh my fellow bald man has to say. Here he is.
One of the things that's going to happen in the ne it's hard to know exactly when it's 10 plus years, but I bet it's not more than 20 years. We're going to start building these giant gigawatt data centers in space. So these giant training clusters, those will be better built in space because we have solar power there 247. uh and it's and and the solar power there is there no clouds and no rain, no weather. So you can build we will be it will we will be able to beat the cost of terrestrial data centers in space in the next couple of decades. And so space will end up being one of the places that keeps making earth better. It already has happened with weather satellites. It's already happened with communication satellites. the next step is going to be data centers and then other kinds of manufacturing.
I cannot tell you how bullish I am about this.
It's um it's an exciting vision. I hope we live to see it. Uh it's going to get a little crowded up there, I think. And of course, you got to get the data back down to Earth, so that'll be a mitigating factor. Uh of course, but as these satellites, low Earth orbit satellites get faster, sure, it's completely possible. I don't think you're going to want um to do this too quickly because it's going to be expensive to put them up there. But as the cost to get into space goes down and maybe the size and the weight of these H100s comes down, you also have the temperature.
And I don't think in this clip we had the temp. He mentioned the temperature, but temperature is the other big issue. It's pretty cold in space as you know. And so you don't have the issue with uh having to cool down these uh you won't have the same cooling issues that you have here. So that's pretty interesting combination. The combination of you know more efficient solar and you don't need to cool it down could be a big win. I'm not sure when this happens but uh there are a number of startups with crazy visions like this and they're worth keeping an eye on.
Yeah. One of them is called StarCloud uh previously known as Lumen Orbit. We actually had them on Twist uh January 17th of this year. If you want to check that out, episode 2073, and the company kind of got us idea for building space-based data center, station Jason, because they were working on doing space-based power, and then they discovered that they were going to lose 95% of that power they generated in space, beaming the power down via microwaves to the Earth. So, they thought, well, all right, if we're going to have power up in space, let's put the data centers up there. They want to use large uh modular 4 km by 4 km solar arrays and to cool it. Back to your point about inspace cooling. They want to use um these they're building their own radiators if you will that are going to use black body radiation to get rid of the excess heat. And the CEO Philip Johnson told me that they need about it's a 1/4 as much space to do the heat dispersion than they need to collect the heat. So you can kind of I think glue it to the back of your solar panels and then have the heat go out that way. A lot of problems with this. Um they think that the biggest risk to this working is launch costs. So essentially this company is a bet that SpaceX's Starship is going to not only get sorted out but do lots and lots of trips up and down. Um one last note from them just based on my interview. There's two markets for this type of work. one is you're dead on getting data to and from the earth, processing in space and sending back down the results. But he also mentioned the CEO of um StarCloud that there's a market in orbit for compute which you can do via lasers between different satellites. So if you're um I don't know Starlink and you want to figure out your next array movement, maybe you just call up an in orbit data center to do that very quickly without any disruption.
I love living in the present, man. It's great. Like we're talking about real data centers in space. So cool. I mean, how do we not have a base on the moon 55 years later? It seems like we should. And if you asked anybody who watched the moon landing, hey, when do you think we'll have a base on the moon? They'd be like, I don't know, 10 years, 5 years. I mean, we're going to go up every year now, right? And it's like, no, it's never going to go back. It has no importance. I I want to know what the hell's going on. I hope I live long enough to understand some nature of the universe.
I'm just curious. I'd like to know YC the video game. This is a find that we had a lot of fun with over here at the Twist offices. Jason, there's a new suite of online games called YC Arena. This is put together by a student. He goes by MHE 100 over on X based in Berlin. Techrunch covered it. I played with one of these games uh in the pre-show period and I have my score. So, I'm kind of curious. Would you like to try to beat me live?
Uh, sure. Absolutely. I I I know exactly how the partners make decisions over there.
I've uh talked to enough of them and uh I know the history of it. Uh it's pretty straightforward and I will uh take this test. Now, this is based on the fact that if you search on YouTube, there's a ton of people just post their Y Cominator applications to YouTube and then they put the URL into the
Yes, there are eight games though, Jason. So, there's a variety of things. I was going to have you do YC guesser, but it sounds like you want to play YC partner Simulator and if you want to play and you're listening in YCA.com. But Jason, pick.
This is the one where they play the video and you say if they got accepted, right?
Uh that's this one right here. Let's give it a try.
Okay, let's do it. All right, let's see. I'm listening in here. YC partner simulator. Start a game.
Hi, little louder. Hi, I'm Arun. We're co-ounders of Va. Uh we've been friends since since 2010 being do being dominates in college. Uh post college, we started this company called analytics which was a driving behavior analytics startup. We were invited to interview for the Y combinator summer 2017 batch.
Okay. Just last month we exited Van analytics to a larger tenatics player entering India.
Okay. They're accepted.
Okay, let's see what you got it right. Wrong call. Rich
um the reason I went with accepted before even hearing their next idea, it's three developers with an exit where I assume they're three developers. They they look like developers or kind of acting like developers or dressed like developers. Um and this it's always at uh Y cominator they want three people they want two to be developers they want multiple co-founders they don't like idea people they don't like pitchy people they like hackers but they're also Indian and depending on what time period YC wasn't investing in India uh all that often and then they started doing more investing in India. Okay so I'm 0 for one. Let's see the next one.
I'm the CEO of Yellow Eggs. And I'm Danielle. I'm the CTO. And our product enables companies to optimize field profitability. We do this by connecting to their CRM system, getting sales and revenue data, and on the other side, connecting to the operational system, getting costs and resource data. Using this data, we analyze each deal to show its expected profitability, the different costs, and how they're allocated. Okay? which enables our customers to optimize the profitability of each deal.
H okay, let's pause this one and oh it has the year there. So this year is 2018 the application. So this is critical.
So in 2018 SAS was hot uh and sales and CRM stuff pretty hot. This seems like a small idea. I think this they're Israeli based on uh the accent. I could be wrong. Um and uh I'm going to say they accepted them. It's technical. It's niche. Let's see.
You folks have any idea how many documents and spreadsheets I got to look at every day, every week, every month. I'm reading this ad offer a document right now. I'm going to need my own personal data center soon. But this is why there's Kod. It's the all-in-one collaborative workspace that's not only helping us here at launch, but working for over 50,000 teams all over the world. And they've got an amazing new feature that we've got to tell you about. It's called Kodrain. It's an AI that's been trained on all of your company's data and information that syncs to all of the tools you're already using. Codabrain is not just going to eliminate constant searching and keep you organized, but it also crunches the numbers and provides you deep research and insights just based on simple prompts. And because it's powered by Grammarly, you know you're going to get clean, clear, crisp responses that are easy for you to understand. So go right now to kod.io/twist io/twist and get six months of the team plan for startups free. That's c oda.io/twist for 6 months free.
Wrong again. So these might be both of these might be too niche the ideas. Um well the first one I didn't even hear the idea but the second one it might have been too niche. Okay let's go next.
Next pitch. Okay here we go.
I'm Jose CEO and co-founder of the most. Hi, I'm Daniel and I'm the CMO and co-founder of The Most. Hello, I'm Antonio Bernardino. I'm co-founder and CTO at The Most. The Most is a news content social media platform where people can uh discover their favorite news and start conversations and events.
No, reject. It's social. No.
There you go. Yeah, definitely not.
So, so what the point here though, Jason, is you can actually see historical YC pitches, see if they got accepted or not, and and learn from them. watching I I got six right, five wrong, so roughly the same ratio that you did.
Uh, but it was fascinating to go back in time, see what was hot then, see how people present themselves, see what worked, and I got to say, it's a great learning experience for founders who want to get in there and work on their pitching. Why not just go see what worked and what did not? I love it.
Yeah, it's um it's fun. You know, there's a there's a lot of like anti-YC vibes on X right now. It seems like there's a younger generation who finds them to be very establishment. Uh, and I think it's because they are very established, right? They've been around now for 20 years. So, it is the establishment. Yeah.
Um, and there's so many people applying, which means more and more people don't get accepted. And their best advice to you is, oh, we'll wait six months and apply again. That's terrible advice. I mean, it's great advice for them. It's terrible advice for you as a founder. Um, if you get rejected, forward your rejection and your application most importantly to ycaunch.co and my team will take a meeting with you and give you some feedback and see if it's right for our program.
I can tell you nobody when you get to a certain number of applications, nobody can tell you the difference between the top 5%. I could tell you like this is a top 10% a top 5% but amongst that 10% I couldn't tell you this one's number one this one's number 10 out of 10 companies you know out of the top 100. You probably would see you know reasonable investors from these programs pick a similar top you know five or six of the top 10 would be the same but they're only accepting 1% Alex so when you're accepting 1% what they're doing is they're just looking for young people without families who are developers. The more developers the better, which means it's naturally going to screw more skew more male. That doesn't mean they're hiring based on gender or they're accepting based on gender. They're accepting based on developer or not. The percentage of female developers, I think, is still under 10% right now is my guess. Uh that could be old data that I have, but it's roughly, you know, in that range. Not computer science graduates. I'm talking about like developers working at a company. And sometimes companies will mix in UX and uh product managers and designers into the development group to kind of make their numbers look better. There's a lot of window dressing that occurs because people are under such uh DEI attacks. Claude says it's about 20%. I think that number is a little inflated, but it could be one in five. What that means is uh they're going to be accepting 80% males. Now, they will look for a female team member when they do their DEI stats because all these programs, all these venture funds were under so much pressure to share their stats. And I think to YC's credit, Tech Stars, they all shared their stats.
And they did. Yeah. And the way they gamed it uh was, is there a female founder on the team? And you know, I'm not I'm not saying gamed it, but just the way they dealt with the criticism that it was overwhelmingly male is to just be like, do you have a female founding team member? And uh you know, you can you you find a lot of them that have, you know, oh, the director of sales, the CMO, the growth hacker, the you know, design as if there are not gender differences in work. Like there are there are definitely careers that draw different groups of people. Doesn't mean that group of people can't do that career. It just means there's a bias that people might pick different careers. The end. Full stop. Uh but that's all over now. In fact, if you were to do this kind of DEI um reporting and stuff like that, you'd probably be advised not to do it and to tell everybody do not pick based on gender, which I did from the beginning. Said don't pick anybody based on gender. Don't pick anybody based on race. Do not pick anybody based on age. Because I knew it was illegal to do that and you would get sued. But you had a counterveailing force which was well why is it so many white Asian Indian guys doing startups and I just said just make it based on merit. The chips will fall they will.
And speaking of uh of things in merit Jason let's talk about launch for a second. You had a tweet that really caught my eye. Um you called launch it's 35 launch patch 35 right now right?
Yeah. Yeah. You called it the best accelerator class in launch's history.
Yeah, which is I mean honestly I mean after 35 classes that's quite the claim. So talk to me about what's going on here at home.
Yeah I just think uh we're getting more applications. Uh I would say since allin started and became like a pop culture phenomenon it's tripled the number of applications and then we're now doing an explicit push to get to 50,000 applications a year. Well most weeks have no problem hitting four or 500 applications but I want to really get it to a thousand a week. Uh now we want a thousand quality applications a week and we do a 100 meetings a week. So we pretty consistently will do a 100 first calls, 15 20 minute calls. The more calls you do, the greater the chance that you're going to find a really outstanding company. That is the nature of what we do in Silicon Valley and in the investing game. So I'm just trying to get to 50,000 applications and then we stay in touch with folks and we have three ways to invest. They can come to Founder University if they're extremely early, if they're in year zero, they're not incorporated or they're recently incorporated, they haven't finished their product. Accelerator if they have finished their product or close to finishing it, maybe have a couple of customers. And then the syndicate, we'll do a direct investment maybe from our fund. Uh and then we have the syndicate, we'll send out highquality companies that are, you know, seed to series A. So we have three ways to invest. And if we can get to uh 200 investments a year next year and have 50,000 applications, that would be a pretty good number, I think, a pretty good balance. And so that's why we're we're saying no to good companies. We're saying no to great companies sometimes, which means if we do accept you, you're great to excellent in in our estimation. And the we try to do is just figure out will a seed fund or you know a venture fund lead a round in this company either in this next round or in two rounds. Do they have that signaling? And I'm I've developed quite an expertise in knowing what the signal is. Product velocity, deep understanding of customers, a kick-ass team that really uh is inspired and hardworking. So when you find those things and then obviously the business model, the go-to market strategy, all those things will also matter, but team, product, customer, that little flywheel always wins and you know, we're we're humble enough to know that we're going to get it wrong most of the time. Uh and uh it's a very humbling hard business to be a venture capitalist. You have these incredible moments where, you know, I wake up and my top two investments of all time, Robin Hood and Uber, if you put the two share prices together, are at $250. And I paid pennies for those shares. So, you can do the math as to like, you know, 67,000x, you know, on Uber and I don't know what I am on Robin Hood. Um, but yeah, it's uh it's humbling because those are two of my first hundred investments. Then there's 400 investments after there. Will those other 400 investments turn into decacorns? Possibly or you know a centurion I guess for a hundred billion dollar company. I I hit $200 billion companies in my first 50 investments. It's kind of crazy.
A trillion dollar company would be a thousand thousand billion. So it be a a mil millicord I guess.
Who knows?
Yeah. Well, centurions, I think, centaillionaires is what they call people author of 100 billion.
So, yeah, versus the pathetic and and and miserly decimillionaires.
They're so annoying about.
I know. They always think they deserve a parking spot. Unbelievable.
Let millionaires. Oh god. Oh god.
So brutal. Uh yeah, don't don't mock the home team over here, Jason.
More work to.
All right. Uh, next up, so I'm really glad you talked about the three things you do, Angel University, the accelerator, and so forth, because um I found a really great question over on the startup subreddit. And I think it's actually great question to answer in this context. So um Jay Jay Furs uh asked, "What do these investors want?" They say Jason they invest as early as possible that they're preede investors but then the complaint was their first question is always what's your MR or monthly recurring revenue and this person put out a really funny kind of like translation guide for venture speak no traction required means you need traction we invest in preede really means we invest in seedstage companies that we're calling preede and then it goes on and on um but you're talking about the companies that you're picking that are of the highest caliber for both Angelu and also for launch accelerator. So what is this person either missing or are not fully understanding about what the VC sees in
um if it's a team that is absolutely spectacular that has worked together before and has had previous success they will invest before the product is launched for everybody else you have to compete for those dollars. So if an investor, let's say a VC fund, you know, $300 million fund with five partners, they each do they can each be on 10 boards, right? Uh and there might be two or three funds in which means net net they probably put 20 to 30 names in each fund. In other words, each partner, if there's five or four of them is uh going to do five or six investments per fund. And then over two or three funds, they might have 12 active board seats or 10 or maybe as high as 15.
So that's 60 board meetings a year, right? if they're quarterly. So, keep that in perspective. They're going to do one deal and if the fund is over three years, that means they're doing a deal a year. Maybe one year they do two, maybe another year they do one, but they're not doing, you know, a deal a month like we're doing a deal a week, two deals a week here because we're an accelerator and an incubator and it's very different. So, you're in competition with everybody else. If you're an unknown group of founders, first-time founders, the VC's probably got a line of people who have between 500,000 in annual revenue and 2 million who can't raise their series A. And those people are really frustrated. So, as frustrated as you are when you hear, "Oh, yeah, no, we'll invest pre-revenue, pre-launch," the caveat to that is if the team is legendary, if the team has had previous exits, if the team, you know, has an incredible skill and they built the most absurd product in the world, maybe. And it means if we can't find anything else, that's not a better bet. So you are taking, you know, what's on the the the website. Hey, we invest preede. They probably have Mhm. Oh, we invest uh when they have traction. Okay. They probably do that the majority of time. But you have to understand who they invest in preede, which is legendary founders with track records.
I'm trying to find the original uh news coverage of Brett Taylor's company Sierra that raised a bunch of money Jason right out of the gate as an example of how to get a lot of money early on even if you don't have a product yet. Brett Taylor founded Friendly which sold to Facebook then he became he was high up at Facebook then he was almost co-CEO at Salesforce and he's done everything right. So if Brett Taylor walks outside and says
I'm starting a company. Checks fall from the sky. That's not everybody. I think that's, you're also obligated, Alex, if you made money with the previous founder's company and they have a new company, you're kind of obligated to say, "Okay, I'll seed the new one."
Oh, because they made you money. They're not expecting you to come back.
Be rude not to. It'd be rude to turn down, you know, if you were RUF and the founder of YouTube started a new company or, you know, you'd be like, "Of course, we'll back it." Pretty hard to say no. Um, and in fact, Sequoia, I don't think invested in Tesla, and they had invested in X and PayPal, right? So, you know, it's pretty good evidence that you should, unless it's a complete vanity project. And even if it is a vanity project, you're probably going to throw in 250 and you might say, "It's not in the fund's thesis, the partners couldn't get around it, you'll personally put in 250k because you made me 10 million, right?" So, yeah.
There's a little bit of that. If your friend stakes you in the poker game, you buy them lunch, you know, come on.
Yeah, something common courtesy. Let's. Yeah. All right. Next up on the docket, an update from our friends over at Figure, the AI robotics company. Jason, we've talked about them quite a lot over the years. Uh, not only because they just raised a $1 billion round at a $39 billion valuation, but also because we talked about their use of their robots inside of a BMW plant. Were they actually live in production? Were they just about to become live in production? We talked about the importance of not BSing your marketing. But, uh, the CEO put out a new video that I'm going to show you of their robot doing the thing we all thought it was going to do. Repetitive work inside of a BMW factory. So, Jason, uh, for the folks out there on the audio version, tell them what's happening.
Uh, well, you've got a robot at a station with doors opening and closing and something being put in front of them for them to manipulate. Doesn't look particularly complicated, but it does look incredibly smooth. And it doesn't look like the robot is tethered, but it obviously is in a controlled box.
Yes.
And the big controversy in the Wall Street Journal story, and I'm not sure when that story was. Um, but maybe it was in 2024. I see there's a story with them, uh, for their funding round, but I'm not sure if that's the BMW story. Um, anyway, this BMW story came out maybe it was in the spring of this year because Figure Robotics.
April 2006. Sorry, April 6, 2025.
So, there this the founder, uh, Brett, is, um, obviously very talented. He has raised a ton of money. And there have been realistic questions of how do you raise $675 million? How to raise a billion without having a product in market to to our previous, uh, conversation. Well, he's gone the non-traditional route in who he's investing from. According to his tweets, he tweeted, you know, he's got a different pool of investors. They're obviously not valuation sensitive. There's been a lot of criticism about SPVs, special purpose vehicles, which we do at the syndicate, which is what Angel List does, and other folks. We do it at an early stage, but these late stage ones, people just take 10% of the money invested. In other words, they don't take carry. So, they're like, "You want to put $10 million into Figure? I'll do that for you. I have an allocation." They go to a bunch of rich people and they $9 million goes into Figure and a million goes into that SPV person's pocket. And maybe the minimum check size is 500. So, they get 20 people and and that's actually maybe a fair trade. Um, if that 10 doubled in value, it'd be 20%, so you take two million. Um, but those SPVs, um, can upset VCs who want to invest in the company at a reasonable price so they can get 10 or 20 times their money, whereas retail investors might be happy to just be in a fun name and tell their friends at the poker game, "I own OpenAI shares, I own SpaceX shares." And if I double my money, I triple my money, and if it's as good as the stock market index or better, I feel pretty good about myself. A VC can't make a living because of all the duds in their portfolio in that way. So, there's a bunch of tension. I think Figure has been caught up in that tension.
Yeah. Where these SPVs and I think VCs are insulted by this concept that there are what they consider huers out there running SPVs and running traffic between high net worth individuals or the dentist crowd, you know, uh, and, you know, taking a vig, you know, a percentage of the investment as opposed to being long-term partners with the company. It feels transactional to them. But founders get to choose, and so do the dentists or the high net worth individuals who, you know, they fancy themselves stock pickers and they want to pick. "I want to be in Figure. It's their right. I want to be in SpaceX. I want to be in OpenAI." We we talked about the $500 million valuation for OpenAI. I could think of many different places to put your money that could get a much better return in public companies, you know, that you can just buy without having to pay a 10% onboarding price or a 5% onboarding price or any carry. You don't have to pay carry when you buy stocks. So, it doesn't make any sense to me. I would I would almost guarantee that if you took, uh, the investment in, I don't know, something like Figure or something like OpenAI, and then you put it into the MAG 7, I don't think you'll see much difference in 10 years.
No. No.
Yeah. So at these high prices because it's entry price matters.
The these low-cost index funds that people now have access to, uh, are such an amazing vehicle for wealth appreciation. I don't think they get enough enough love. Uh, but Jason, Brett got back to you. You asked him very specifically what's going on in this video, and Brett responded, uh, just a few minutes ago, actually. Uh, this is the body shop production line. Metal parts are picked up and loaded into a fixture that then gets welded together on the other side. So instead of having, um, you know, Eminem and 8 Mile standing there with the stamping machine, it's going to be our Figure robot. And, you know, in his original tweet, he said they're doing this in 10-hour shifts.
Not bad.
Yeah. I mean, that 10-hour shift is probably predicated on a human monitoring it for now. And if they can do the simplest part of the task, which is aligning the pieces, that seems pretty reasonable. And then the other person has to do the difficult task, the welding of the pieces. Um, that means they're only a couple of years from this, uh, robot doing the welding as well. Uh, and so.
Can Figure afford to have that still be a couple of years out? Because I'm bullish, but also I'm long-term bullish, short-term uncertain about these companies.
I think if you raise a billion dollars and you're burning a hundred million a year, you have 10 years of runway. So, they can literally burn $10 million a month for a hundred months.
A billion dollars.
Let that sink in. 10 billion is a lot of money, folks.
Wow.
Yes.
All right. I mean, it's almost. Yeah. I mean, it's just it's or if they if that money is just going to last them a year, let's say they were just going to they were going to spend that money over a year, they'd be spend they could spend three million, two, $2.73, $3 million per day. That's a lot of money. So, I think that they have plenty of runway and so they're in no rush. Thanks for Brett, uh, for getting back to us. Appreciate that.
Yeah, come on the show anytime.
We would love Brett, we'd love to have you on. We have so many questions about robots. Um, other companies in the space, of course, Tesla's Optimus, Agility Robotics, Boston Dynamics, Unitry, Appronic, and the list. Jason, goes on and on and on. Billions of dollars invested. Space to watch. All right.
I think Figure will be one of the top three. I think Optimus will be number one. Figure will be in the top three, four, five, and then you got a lot of people in China working on this stuff. So.
Yeah.
There'll be a lot of low course, low-cost folks in China doing this.
All right. Uh, next up on the docket, Jason, the benefits of onshoring. You flagged an article from our friends over at the Wall Street Journal talking about how Sharpies, the pen that everyone likes, uh, found a way to make their actual items more cheaply in the US, which was a surprise to me. What, uh, what here caught your attention?
Um, just the headline. Uh, I thought it was kind of funny that, you know, uh, this is the least important product in the world to onshore. Um, but the fact that it was able to be done cheaper in the US just led me to wonder why. Because uh the only cost difference I can think of is the shipping costs. And the shipping cost for pens doesn't, you know, it's not like a very heavy object, but it's a low margin. It's got to be a low margin business. And so I wondered if this was cap or not, and if this was being done as Trump bait, uh, because Trump wants to on President Trump, uh, wants to onboard, uh, reshore and build factories here. I could see people doing it to curry favor with the administration, but I would think that this had to happen before.
Oh yeah, this is a much.
Long before if they've actually completed it. But, um, you know, it's, uh, also the company seems to be a bit of a disaster. Their, um, their balance sheet, people were commenting on that. But what's the secret here? Why are they able to do this cheaper here?
So, uh, a lot of work. So, they started this project back in 2018. This is a facility in Mville, Tennessee, a town that I'm not going to lie, until this article, I had never heard of, Jason. Um, but they've done a couple things. One is they've invested in robotics, and they've also invested in upskilling their staff. And the combination of the two has yielded a workforce that has lower turnover and higher efficiency. And so, they're able to make products at a cost-effective basis here in the United States. Um, I think.
So, robots is the answer.
Robots and more skilled humans. Uh, they did have a a program to reduce turnover rate for manual jobs and, um, so I think they they tried to reskill their staff so that way they could support the factory versus just getting rid of the people. We've talked quite a lot about mentorship and, um, upskilling people here on the show in the last 18 months as AI's come onto the scene. Uh, but I think that the thing that I want to show is just this picture from the Journal. Um, sorry.
And I love a good Sharpie. I'll be totally honest. I'm a Sharpie fan. I love the brand. I would never buy any other permanent marker because why would you? I'd trust a Sharpie. It's like a consistent brand promise. Sharpie, it's like Gorilla Glue.
Like I was ordering some double-sided tape from Gorilla Glue, you know, and like that that really industrial stuff. And I'm like, why would I ever trust the Amazon choice when Gorilla Glue is available or Sharpies available? I don't want to buy the Amazon basic. I want to buy the real deal.
Yeah. The Sharpies are such a tremendous product. Um, but this picture here, Jason, from the Wall Street Journal, I think shows exactly what you're saying. This is a this shows a woman working on a robotic.
I'm going to guess, uh, incline of some sort here. And it's it's highly technical, but it does show that you can with foresight, time, and investment, you can make stuff here at a cost-effective basis. I think the difference is that it's hard. Like I bet this was not easy. It's probably very easy to get manufacturing spun up in China where there's a lower cost of labor and they already have more industrial machines set up for this type of work. But you you can do it. It's a will question, not a how question, I think.
Yeah. And if you think about people graduating from college or entrepreneurs, if I told you you can design a product, have somebody else make it, and then you get to sell it and take 95% of the profits. And the person who made it is like marking it up 5% from their costs. Which business would you want to be in?
You obviously pick the former, not the latter. And Steve Jobs had this same realization. It used to be his machines used to be made in Cupertino, then they were kind of assembled in Cupertino, and now they're designed in Cupertino. Foxconn, I think they're like a cost plus 5% manufacturer. What does Foxconn make on the average phone? Like if the and versus what Apple makes?
Apple. And I was I literally was explaining this to my 15-year-old about how even though the phones are made in China or India or wherever, all the profits reside here in America. And so American workers like turning tiny screws and, you know, doing six days a week, 10-hour shifts, it's just it's not going to happen, um, all that often.
Oh man, Jason, this is brutal. So I'm on, um, so Foxconn, I'm on their latest, uh, monthly earnings report. Just going to quote here, gross profit margin, operating profit margin, and net profit margin were 6.3%, 3.2%, and 2.5% respectively.
Which is exactly what I said. They make 5% something like that.
That's almost groceries. That's almost groceries for some of the hardest work in the world. Well, because if they don't make it, they know they can make it in India or they can make it in Vietnam. And if the factory in Shenzhen can't make it, they can go to the one in Guangzhou. You know, it's like there's massive competition for this labor. And you might be talking about making 10 million phones. In which case, if you're making 10 million phones and you're making $5 on each.
Uh, yeah, it's a lot of money. I mean, if you're making a hundred million, now you got a half billion dollars in revenue. It's good money. Um, and if you've got more labor than you could ever use, it's great money.
Ah, but that's changing. That's the interesting thing about the Chinese workforce is they're they're past peak working age population. Similar things we've seen in Japan and so forth. So, I think when we think about the the great investment and acceleration of robotics in Chinese manufacturing, it's not only making things more efficient, it's also, I think, fending off just having fewer bodies to throw up problems.
Um, absolutely. Let's do the Deloitte AI slop story because we're running out of time here and this one, when I saw you guys talking about it, I was like, oh my lord.
All right, so Deloitte, of course, is a well-known global consulting group. If you know McKinsey, you know Deloitte. They are a private company, so you can't go look up their stock price. But they did get in trouble over in Australia because they got caught using generative AI, Jason, to put together a report that they charged, uh, in Australian dollars, $440K for. Classic generative AI mistakes, citing sources that weren't there, all sorts of things. But the real the real killing blow, if you will, came from the, uh, labor senator Deborah O'Neill. She says that Deloitte has a human intelligence problem. The situation would be quote laughable if it wasn't so lamentable. And then she said, perhaps instead of a big consulting firm, procurers would be better off signing up for a ChatGPT subscription.
I mean, that that's what consultants are afraid of being replaced by an AI chatbot.
Yeah. Now, what you're going to have to do, I I don't think that these firms are going to go away, um, like Deloitte. I think what they're going to do is they will start and they will have these tools, you know, deep research tools at their fingertips.
It'll present them a bunch of information. Then their job is going to be doing first-person interviews, acts of actual research, going to a factory, interviewing customers, all that stuff. Because if all the world's knowledge isn't just available online and your job as a consultant was to collate that information and make sense of it, well, if the AI can collate, not not just collate it, but they can also make sense of it.
Great. If they do a reasonable enough job, it's your job to check it and then start your work from that point. Not just hit publish or send and then go off and, oh, sorry, baked that out. But then go off and, you know, I don't know, do like these video, the what we see in these day of the life of videos where people are like, "Then I got a matcha, and then of course I went to yoga, and then of course I went to the salad bar, and then I met with my friends and we played foosball." You ever see these videos that they're on?
I mean, I I did it in 2021. I didn't know they were still making them. They're back. And the second I see one of those, you post one of those working for me, it's instant firing. Instant. Like somebody make a day in the life working for me and it's like, you see somebody schedule it's like, "I met with four founders, uh, then I ate lunch at my desk, then I did a meeting with seven of my teammates to introduce founders to VCs to see if they could help close their round, then I did four due diligence packages and then I collapsed at my desk and then."
And then I died. And then I didn't even. Absolutely. Like I mean, come on. Like the matcha bar is there for you to get to. You want to get a matcha because you got customer partner comes by the office and you want to hang out outside. That's totally fine, but get the matcha. Like let's get back to work, you know, take a little walk for 15 minutes, but like it doesn't need to be like you could do yoga after work. I'm not against yoga, but what is this? Yoga at work.
Yeah. Like what world do you think this is? Google in 2007? Like like calm that. Everybody also also pro tip. If you have a great life and you're getting paid a lot of money to not do a lot of work, shut up.
Like like why why would you tell on yourself? I've never had one of those jobs, but I would. And if I if I ever get one.
Man, imagine we got that job. We do one podcast a week, but then we walk around and drink matcha lattes and get toasted bagels.
Do you know what I would say? I would say in that situation, I would tell everybody, "Oh my gosh, I have never had more work to do. I am drowning."
Yes. Yes.
Yes.
Anyway, um.
Uh, don't don't cheat. Come on.
I mean, it feels like the the this feels like cheating. That's the problem. And this is where I think you're going to have to show your work going forward. And it's going to have to happen when you're in high school and college all the way to when you're, you know, if if you can get a job at like a place like McKenzie or Deloitte, you're going to have to show your work.
And the way that's going to work is in a document. They now have, uh, software in colleges where you write your paper and it doesn't let you cut and paste into the document. They've removed cutting and pasting. You have to type.
And you have to form your own sentences, which I taught journalists. I had journalists who were cutting and pasting other people's work, pasting it into the document and then editing a bit, and they would get caught. And it'd be like, "Are you guys dumb?" Like literally just if you read in the Wall Street Journal that BMW had the Figure robot and they were testing it, but they wasn't in production and then Brett said this on whatever, just say it in your own words. And if you don't know how to say it in your own words, you can just hit the te the speech to text button now, say it out loud, watch the text go by, and then clean it up.
Yeah. I mean, it's so I don't want to use the R word on air, but it is so R-word. What's going on here with people that you've been given this gift of a tool, uh, an LLM that can answer any question, and then you decide, "Oh, that's how I'm going to do my job. I'm being paid money for it." No, if you can get that information in four minutes, the other 56 minutes of the hour is you improving it.
Yes.
Find something to do, people. I I now get to use ChatGPT for a lot of small things. And you know what that makes me? I'm able to do more things.
All right. Um.
A better show, more facts, more answers. Absolutely.
Instead of us saying, "Oh, we'll we'll get you that answer for the next show." It's like, "No, we're going to get it this show. We're going to check the source."
Yes. But Jason, speaking of founder, speaking of use of AI tools, why don't we hear from a founder that used Gamma to create an awesome pitch deck that they want to show to you?
Oh, I would love to do that. Is this our Twist pitch deck competition brought to you by Gamma? It is. It is our first pitch deck competition brought to you by.
Gamma. Yes, we're gonna we're gonna take pitches from 10 amazing startups and, um, then, uh, I think we're going to invest in one of them. Yeah.
I believe that's the plan. Today we're going to have Ryan Yanelli from Next Visit AI. Ryan.
Hi Alex. Hi, Jason.
How are you?
Oh, you're in the accelerator, I see.
Yes, I am.
Beautiful. Oh, I could see by your background. Okay, so, uh, we'll get to hear your pitch. Uh, you ready to go? Share your screen.
Absolutely. And we'll start your pitch in.
Sure. All righty. Here we go. In. Sergio, tell me if it's okay. Now, uh, before we get started, Brian, you went to Foundry University in our 10th, uh, cohort, and then you're in Launch Accelerator's 35th cohort.
Yeah.
Absolutely. Yes. Amazing. Uh, so when you came to Founder University, where were you at as an entrepreneur, as a company?
Um, I mean, after going through Founders University, now being in the accelerator, looking back at it, it wasn't even a company before. Um, it became a company once we actually followed everything that we were taught in Founders University.
Yeah.
So, just invaluable for us.
Oh, great. So, lots of blocking and tackling, incorporating, doing a cap table, accounting, all that important stuff. Yeah, you learn some of those, uh, tactical things and then getting a product to market, obviously.
Absolutely. Yeah. Hi, I'm Ryan Enelli, CTO and co-founder of Next Visit AI. We saw burnout by doing the charting so doctors can do the healing. I spent years going to doctors seeking answers and ended up hours away from my death because my care was fragmented. My providers were overloaded with paperwork. My history was scattered and it resulted in my care being neglected. I'm not alone. One in four patient charts contain errors. Clinicians spend over three hours a day on charting, and this leads to burnout. I want you to meet Dr. Rathor. Before Next Visit, he saw 16 patients a day, was burnt out, and had clinical errors. Now he sees 24 patients a day, saves time, and also saw a 30% revenue increase. Here's how it works. Dr. Rather selects a patient, starts his session, and Next Visit listens. Clinical data is built in real time with deep insights into the patient chart. When the patient leaves, the chart is finished and the notes reviewed by Dr. Rathor. Then it's ready for billing. It's fast, EHR ready, and HIPAA compliant. Since launch, we've gained 311 users and have 68 paying customers. And our customers are addicted. We have 1.6% 6% churn, 24% conversion, and a near-perfect NPS score. We've scaled to $9,000 MRR since launch. Our CAC is $189 with a $1,700 LTV, and our average revenue per user is $133 per month. We're starting with behavioral health in the US. A $2 billion TAM capturing 5% or 60,000 customers gets us to $100 million ARR. Most competitors are just scribes. We're a complete platform that providers trust. We provide real-time clinical decision support, build accurate data, and become irreplaceable. I'm a full stack engineer with 15 years of experience in enterprise environments. My co-founder, Dr. Rafi, is a psychiatrist with over 15 years of delivering patient care. We're Next Visit AI. We solve burnout by doing the charting so doctors can do the healing. Thank you.
Unbelievable. Incredible. I'll give a little golf clap here. Get a little golf clap going. That was perfect. A perfect pitch. You explained exactly what the problem was. You explained what the solution is and the opportunity in terms of the total addressable market and why you are uniquely and your partner who's a psychiatrist are uniquely qualified to do this. Uh, so this is as close to a perfect pitch as you can get. If I were to score it, maybe 8.5 out of 10. I don't give 10. So, you know, 8.5, 9, and 9.5 would be the three choices. I think making sure people understand this is for psychiatrists and psychiatry and that you're very focused on that because listen, it's only a two-minute pitch, what we gave you. So, you can't get to everything in a longer pitch. Obviously, you would get into competition and the niche, but as close as you can get to a perfect pitch. And I do wonder if you're charging too little, but at $133 a month is what you're averaging. Uh, Alex, that's, you know, what are we talking about here? We're talking about almost $1,500, uh, per year per doctor.
They use it for five years. $7,500.
Those doctors make a half million dollars a year. So, you're only char you're not even charging 1% of their total billable value. They char what do they what do they bill out a year? A million dollars, two million. I know from just our statistics that they can bill an extra probably $1,500 a day once they start using us, which is.
You have to charge more. My wife is a practicing psychiatrist.
Oh.
Bro, Brian, you could like I I was on your site before, so I looked into this. This is what I was going to say, Jason, is $199 a month for unlimited notes for your provider plus plan is just too small. Not only can psychiatrists afford it themselves if they're in private practice, but if you tell a hospital they can get 30% more patients through a one of their rare hard-to-get providers, especially if they're a child psychiatrist, man, that's worth two zeros more per month than your.
Yeah. And I think the way to frame it is, um, this is either another professional doing this. So this is your, uh, billing notetaker. It's like a it's a it's a salaried person. So it's $20,000 a year equivalent or something. Or to do it per record or to do it per billing. And so there's consumption-based pricing, which you might want to experiment with at some time, which is, "Hey, every time you do one of these, it's $10." And it's like, what is an average visit? $500. So that's not even or $1,000. It's not even 1% of a $1,000 visit. So it might be that consumption-based pricing would be better here. Either way, if you're undercharging for it now, that means, uh, you know, that's why your conversion rate is so high. What are you converting right now, Ryan, in terms of, um, from a demo to a purchase?
It's a very, you know, it fluctuates a lot right now because last month we had a a huge uptick in users, probably 300 new users. Um, and what we're seeing is like now just this week they're starting to actually start paying for it.
Ah.
So.
So it's hard. Yeah, it's hard to know early in a startup. So this is where you're, this is like a next card you have to turn over. It was a poker game. You know, you've seen the flop, but the turn is going to tell you like your conversion data. You get a lot more information on that turn card. Um, because 80% of the information is now on the board. There's only 20% left to come. That's going to be really important for you to understand is, you know, of this source, could be a conference, could be a referral, uh, could be an insurance, uh, an insurance partner asking people to use it, whatever it is, could be from a paid ad. Of those, how many actually take a demo? Of people who complete a demo, how many actually convert to paid? And then how do you convert them to paid? So I would say you do the demo with them and then at the end of it say, "Can I have your credit card and sign you up right now? We'd love to sign you up right now. Can can I get your credit card number? I'll put it in. It's only $150 a month. You can cancel at any time." And that's the then you might see your conversions if you have the right people doing the onboarding. Oo, like founder onboarding. This is what Raul did. Now, Raul was a very successful, uh, executive, uh, at Superhuman. He did these himself this way as a founder. He he got that knowledge and when he onboarded me as one of the first users and the first investor, he said, "Okay, and now that you've seen it, I'll take you. Would you like me to sign you up?" I said, "Yeah, of course I am." He signs me up. He goes, and he gets to the screen. He goes, "Okay, now it's time for your credit card. Can I have a credit card?" I'm at Cyclass Coffee. I'm like, "I just put $500,000 into this and you're going to make me pay 30 bucks a month." Absolutely. I gave him the card. I was like, "Absolutely. I want to pay for it." So, I think they're going to absolutely want to pay. Um, since this is so perfect, I don't have too much feedback from you for you. You're also for a CTO, you're a gifted natural presenter. Did you ever do a little Summer Stock or something? Were you in the poetry slam contest? What's the story, Ryan?
No. Uh, I actually wasn't sure I was going to do on this because I just gave presentations in high school and that was really it. But, uh, no, I just, uh, really enjoy what we do. Like I, you know, I'll get an idea at 3:00 AM. I'll wake up and I can't fall back asleep until I do it. So.
Love it. I'm just very.
You know, we call that we call that founder product fit or founder market fit. You really love the market and you're you're engaged and that comes through in a presentation. You can tell if somebody's faking it and going through the motions, it's obvious you're not. So, even though you weren't trained to be on stage, man, your enthusiasm came through, your excitement, your passion for this product came through. You're going to do fantastic with this business. I think this is going to be a tremendous success. Uh, and I'm really excited that you allowed us to participate in it. Brian, uh, any questions for us? Any way we can be helpful or any questions about, uh, how the business is going?
I mean, I would just ask if you were me, you know, what would you just obsessively focus on right now?
Okay. Uh, product and customer are probably the most important thing right now so that you have, you want to have very low churn. Uh, because keeping a customer that you've already sold is a lot less work than finding another customer and onboarding them, as you probably learned. So checking in with the people who aren't using it, figuring out why, figuring out, "Oh, do they need an onboarding in person or are they fine doing it over Zoom?" You know, there could be all kinds of different factors, but really studying the churn data and the engagement data, you can't go wrong. So look at that engagement data. How are they using it? How often are they using it? Did they do 10 meetings today and not use it? What happened? Was it they didn't like having their laptop out? Okay, maybe they need an iPad. Maybe they need an iPad mini is the best tool for this. Like there could be all kinds of factors. Maybe like their desk, it, you know, like turning their monitor around on their desk is uncomfortable or whatever and they don't like that vibe when they're meeting with the patient. So you get them a small computer with a small LCD portable monitor and a Mac mini underneath. Maybe that form factor works better. So there's going to be little nuances like that. You can never go wrong studying the engagement data and watching a customer use your product. So that's where I would spend your time. Alex, any thoughts there?
I think that's dead on. I'm just really curious to see what they can do with AI on the, uh, on the note-taking side because doctors spend too much time doing it and the resulting product is trash. So if they can make it not only faster, more efficient, but also better, that's going to make them, I think, an unstoppable market force because doctors hate notes.
Famously, just they do them all the time and they hate them.
Absolutely. It is the bane of their existence. It's like for sales people, like updating the CRM, they're like, "Ah, I'm a good salesperson. Why do I have to do this?" It's like, "I'm a good psychiatrist. Why do I have to do this?" You know, it's like, "Well, because you want to get paid and."
Yeah, because billing, that's.
Billing is kind of important. All right, man. Great job, Brian. And, uh, you can find out more at nextvisit.ai.
Just so great when we have, uh, one of our founders come on because they've been in pitch training and you can learn more about Gamma.
Where what how can they learn more about Gamma? If you're feeling inspired, my friends, head to gamma.app and start working on an idea of your own, and maybe we'll see you here on a future episode of This Week in Startups.
Do it. Mondays and Wednesdays and Fridays, Jason. There's always more TWIST on the way.
Awesome. All right, everybody. Speaking of that, we'll see you on Is there a show tomorrow, uh, Lon, or is it Monday, Wednesday? Okay, back on Wednesday. We'll see everybody on Wednesday, October 8th. Go to YouTube.com, search for This Week in Startups, and click subscribe, hit the bell, get all the notifications so you can talk to us live. We'll see you next time. Thanks, Alex.
Thanks, Jason. Bye everybody.