0:00I think humans will be like a biological vessel for like AI. Dying will be a thing that maybe isn't like real in like a few years from now. Joining me in the horse seat is Kush. He's the co-founder and CEO of OR. If you don't know, OR is the compute market building price indices and futures for GPU compute.
0:17I think infrastructure investors actually like end up winning because they will always keep collecting rent.
0:22This is an incredible analysis and breakdown on the cost of intelligence today, the future cost of chips, the future cost of compute, everything that we don't know that we need to know as we look forward. The easy categories are all done. Do you miss out by not being in the valley?
0:36No, I think New York's like much better. We have this whole theory that people in the valley are like coin operated. I think compute is a larger market than oil will be or any of the other commodities have been for the past like 100 years, 150 years. this and so much more in what is an incredible discussion with Kush. Ready to go.
1:06Kush. Dude, I'm so excited for this. So, people don't know, but I got a text from Miles at Thrive being like, "You have to meet Kush. He is one of the most exceptional founders." And I think Miles is probably one of the smartest guys in Venture alongside Pat Grady. And so when he said that, I was like, "Fucking introduced me." And then we spoke and I was like, "Wow, he is as good as Miles said." And so I'm so thrilled that you would join me in the studio today, dude.
1:29No, thank you for having me. It's exciting to be here.
1:31Now, I want to start with a little bit of texture on you before we just get into the meat. We've got many kind of weird and quirky things about you. With greatest respect, um, one that I want to start is you created a viral sports betting platform in college.
1:44Um, so Wayne and I really wanted to intern at like this company called Novig. at the Jacob.
1:50Yeah, we were like freshman or like sophomores um at MIT and they like didn't let us basically intern there.
1:55Um and so we were basically like all right, you know, [ __ ] you. Like we're just going to go like do this ourselves. U which like at at the time like we were only good at like coding. We were really bad at like everything else. Like we were really bad at commercials, legal.
2:08We had no idea how hard it was. We were just super good at like coding. So I ended up building it was just like an app that you could like buy and sell. Like it was just like sport. It's what Novig and Khy and like Polymark and all the apps are today where you can like buy into sort of like sports things and it was like you can bet on player stats.
2:24Um and ended up just like working with all the fraternities at MIT. So you could literally go and like buy and sell um different sort of like players or odds. U but it was cool. It it never really like took off past like MIT but within like the school it was pretty sick.
2:39Did you tell Jacob at Novig this?
2:41No, I' I've not told him that.
2:44I don't think he even heard about it because it was so like we were so bad at like go to market and all the other stuff besides like building the app that it just never worked out.
2:53That is hilarious. So, I'm going to send him this clip right afterwards. Um, another kind of uh inspirational person that you very much look up to is is Travis. I heard this from some of your investors and actually some of your teammates. Can you talk to me about what would TK do? Uh I so I I I'm a huge like Travis Kalanick fan and I the funniest thing is like a fortune reporter asked me like what do you think about Travis?
3:15Um and like our media team was like really mad because I just said oh he's the Steve Jobs of my generation. U but I but I really think like there's so much to like learn from like older entrepreneurs that have like done this before and I'm like always like oh like if if we're in this situation like what is like Travis going to do here?
3:32What is the result of that? Is it be more aggressive? Is it be more ambitious? way way more like aggressive and ambitious. Like I think a a part of me that's like I'm really bad at is like not being ambitious and not being aggressive enough. I'm like a very nice person. Like I will not like um fight back or like push back on things and when I need to. And so like I think Travis is the exact opposite of that.
3:53And so it's like it's a good thing for me to learn. What were you not aggressive enough on that with hindsight you're like damn I should have been more so many of our like early commercial agreements like we sold a lot of our data um early on to large like hedge funds and like trading shops market those guys are like sharks like their whole job is like I'm I'm just going to make as much money as possible and we're pretty bad so we end up selling it for like 5,000 and like 10,000 a month and it was like I was like are we like losing here cuz they would immediately say yes. Um, then I realized like they're willing to pay like $100,000 a
4:23month. They're like 500k a month. And I'm like, "Oh, wow." Like we are like underpricing this crazily.
4:30That is amazing. You almost sound European with some of your like I'm not pushing hard enough. I'm not aggressive enough. Welcome to London, baby. Um, you are quite nuts though in the way that you did a marathon with no training. And several of your investors told me about this.
4:46Yeah. So, so it's it's funny. So, I I used to like I did a bunch of running in like college. Um I was never that good, but I like I ran a half marathon because like my friends told me I could never run like more than like a mile. So, I ran a half marathon. That ended up being like I ran like a marathon um while still in school. And then I signed up for another marathon and I didn't know I was going to start the company. Um so, I signed up for it in July and we started the company in September. And then, as you probably realize, like when you start a startup, like this is like your whole life. Like there's no time to do
5:14other stuff. In marathon training, it takes like two hours a day. You like go run for an hour and then you have to like shower and like put on your shoes and all these things. It takes a time off your day and like I just didn't have two hours. So, I literally like didn't have time to train for the marathon. And day comes like I'm like all right, we have to do it. It's like there's no like I'm not a quitter. Like we're not like we're not going to like not do this. And that was like the time when Jensen like released like it was a podcast when he was like we not a car.
5:42Oh yeah. Um, so I woke up in the morning like we are not a car. Like we're going to go run this like u marathon and I ended up just like I mean I did it. It was very bad. Like it was painful but I think my time was not that bad either.
5:53It was like um 3 hours and like 30 minutes or something. Um so it was like decent but it was terrible experience. I trained for a long time and I still did it in like 3 hours 45. So um dude I absolutely love that. Well I I think that's sufficient context. When we were chatting before, there were elements that I picked up on that I think I really want to dive into straight away, which is number one statement, AI can win while
6:21infrastructure investors lose. Can you unpack like what do you mean by that?
6:26Yeah, I don't think like infrastructure Okay, one is like I don't think infrastructure investors will ever lose.
6:30The way we think about it is like infrastructure investors are like landlords um for like an apartment or rent. So the easy analogy is to use is like data centers are just like apartment buildings where someone buys the apartment building or and then the actual people that take the ownership of the GPUs are renting the GPUs from the apartment building or like they're renting units within the apartment building. So that's companies like Fireworks, Modal Fal, B 10, um all the inference providers along with like OpenAI anthropic depending how the deals
6:59are structured. Um, and so you could argue that like I think infrastructure investors actually like end up winning because they will always keep collecting rent as long as the tenants there's tenants to live in the apartment building. out. If it turns out the city's like doing bad, which is like in this case the city is like AI, then that means like not a lot of people will like want to be renting an apartment um or in the case that like there's not enough demand for apartments and everyone's kind of looking to buy a house quote unquote is like I'm going to go build my
7:29own data center. Um and so there's all these different sort of ways, but I I don't think at the end of the day like infrastructure investors like end up like losing. I I think there would be a lot of larger enterprises who say there is still a remaining ROI question for sure and there is still a cost question and if there are cracks in enterprise demand that would trickle down to a lot of the providers and that would trickle down to the infrastructure investors.
7:51Yeah, I I think so for sure. I think the infrastructure investors it's it's again it's the same way as like rent works right if there's like the economy is doing bad obviously like prices for rent are going to fall down and that means the cost of like GPU hour will fall down because there's not enough demand for the GPUs and it's really just like it is the end of the day like a supply and demand game where demand if enterprise demand continues if it goes down obviously like supply of the compute will go up because there's just not enough demand for the tokens or the training or whatever that may be. How
8:20undersupplied are we given the lead time for these projects being?
8:23It it takes like I mean it's very undersupplied. Like if you want to go to coreweave and buy something, it's probably like you're probably quoted like Q4, 2027, Q1, and 2028 at this point. Um if you want to do like a build your own and get chips from the OEM, the OEMs usually like lie to you and will tell you that like chips are going to come earlier than they are, but it's probably like 20 22 weeks time period to actually get the chips and then install takes like another like 2 to 3 weeks.
8:48Um, and this is on like a small scale. I'm talking like half a year to build your own.
8:52Are you concerned by the debt cycle that we're seeing to fuel this infrastructure?
8:57I don't think really because today we've seen that actual demand for tokens and the output of the AI producers actually like ends up being useful. I think the public markets are way too scared about oh AI is like not useful. I'm like have you used chatbt? Have you used like any of the products? Like clearly there's like a use case for this technology and which means if like there is a use case for this technology the thing that powers this is like the data centers the chips all these things like there will be uses of this like these chips for a long time past the period and like
9:26prices have also gone down gone up over time like the prices for literally H100s and even like the B200s or B300s are way higher than when they started.
9:35I think Gavin Baker said every single day for the last like 60 days they have increased in price.
9:40Yes. Yeah. Yeah, I think literally chip prices have been increasing and that's because the demand for AI is just like continuing to increase and that demand is like exponential and you probably seen all the revenue curves for all the companies are just like straight up. Um, and that's and that's literally because like what is what do these companies do?
9:56They sell tokens. That means in order to sell the tokens, you'd either need to get like way more efficient at like generating the tokens or just gen like get way more compute. And you're going to have to do a mix of like both things. And part of it is like if I want to get way more compute, I'm just going to have to start paying more than like everyone else.
10:13What does no one see about chip pricing that you see because of your very unique position?
10:18I think the biggest thing is like everyone's I don't know, maybe everyone does see that it's just like going up over for a long time. But when we started the company, a lot of people thought like, oh, the curve for like the forward curve for chip prices is actually like backwardated in the sense that like the 3 years out from now or 5 years out from now, the price for GP hour is going to be less. And so many like agreements you would do early on was like people would offer like free call options to people that would buy the chips where they'd say, "Okay, three years from now if I want to renew the
10:47contract for another two years, I'll just put in like I can renew it for like $2.50." Like no one's going to renew that. And now you look at it and someone asks you, oh, can I put this like call option in? So 3 years from now I want to renew for like the same price that it is today. No one gives that for free because I think people of in the market see that oh maybe like three years from now the trip price are actually like higher than they are today. I would not be surprised if their Reuben costs are like 20 $30 per GPU hour.
11:15I think it's like seven or eight for like a 5year. I think I genuinely think like the demand is like insane. It's a bad thing and like a sort of like a good thing that is such this is pure just supply and demand just like markets. So why does everyone [ __ ] on these neo cloud providers like Nabius and Coreweave and why is their debt like bonds being through the roof?
11:35I mean I think it's because they're levered in the they're levered to the point that like they believe that the residual value of the chips and the residual value of like AI will continue to progress over time. Um and people just think like oh like this is a bubble this is not real and that that kind of like stacks on to and onto and onto each other. But if you actually look at the data, the price per GPU hour is still increasing. So I like don't understand why like people are like oh we need to go like short like coreweave or NBS part
12:04of it also like a credit question and that like the tenants of coreweave like will they be able to pay their rent for the GPU hour and that that in itself is like it's an AI question because the tenants are like AI companies.
12:17Sure. This is your fowls and your base hands.
12:19Yeah. Like it's like will they be able to continue to pay the rent for coreweave over time? Totally get that. So you would buy Core We I would go long Coranius. Yeah. Yeah. I'm not allowed to uh to publicly invest. Um but I would if I could I would go like very long.
12:38So I love it when you have like a very clear we believe that AI demand will continue. We believe that there'll be a series of knock on. Yeah. What's the bare case? We've got the bull case.
12:46What's the bare case? Uh I think the bare case is that like AI demand doesn't continue like maybe like in the other option is like okay people stop buying tokens enterprises don't see that AI is actually impacting their business or helping their business which I think is also partly true today like if you look at like all these like studies from the consulting firms I used to work at BCG um so I'll like quote their their like study you used to work at BCG I hated it yeah it was it was that's why I started the company partly um cuz I
13:14didn't want to work a corporate job but I I think partly is like but if you look at like all the studies it's like oh have the enterprises actually like implemented AI like are they like being like useful with AI and I think the answer so far has been like not yet um I don't think it's been very useful for many of the large companies so far and maybe that's because it's like implementation problem and all these things and like I think demand is still increasing so it's not a problem but let's say it doesn't work out maybe Coca-Cola or like McDonald's or many like the large Fortune 500 companies
13:43realize like oh AI doesn't really change my business um for the most part and in that case it's like okay then all these like chips and all this computer I bought and like you're kind of over buying you have to buy before you like know there's like demand for it um doesn't work out and then the people can't pay the rent and then um you start defaulting on these chips and then creditors basically have these chips that are worth like what like they're worth like the metals in them or they're worth something in them that is not the
14:12actual value that you can like plug it in and use it for. How do you advise founders on forecasting the buy of compute? Like I asked this from Eric at my Yeah, I watched the So Eric what Eric said he buys like 18 to like 20 uh weeks in advance. I think it he should start behaving he's probably going to start buying like 22 to like 26 weeks now just lead times increased for everything. Um but that's really like the case. So part of like the company for us is like to help solve this problem. It's like okay
14:41people need to buy before they like know what's going to happen. So that needs you need to hedge your risk somehow, right? There's a there needs to be a somehow way like if I'm buying today, I don't actually know what prices are 22 26 weeks from now. Maybe they're higher, maybe they're lower. Um, who knows? It's a long time. But I would genuinely say like if you if people have the cash and many of these like Fortune 500 companies especially if they're like, "Oh, I'm going to invest in AI and I have the cash, just go buy the chips, go build like the data centers or go do something
15:10because I think prices will continue to increase and you're just going to be like left paying more and more every single time. And let's say you don't use them, then you can just do exactly what Elon does. We look, we allow you to rent the chips through us, but even even if we don't want to go through us, do exactly what Elon does. He goes to like SpaceX and he says, "Oh, I have extra chips." Anthropic, pay this much. Hey, pay this much. Um, and so you can just start becoming like a Neocloud and start charging like as much as you want for the chips because you have the money and the credit for them.
15:39Elon has such a good business being in Neocloud. It amazes me he spends so much time building and everything else.
15:46Um, I mean, look, I I think the I think their models are pretty good, too. We're I mean, look, we're big users of like Cursor and Grockbot and all the applications. Yeah, our team I I I'm not like so I'll I'll walk you through our stack. Okay, so we use Grockbot for like all the daily tasks like I don't have an assistant or any of these things. Just pure Grockbot. Um and it works well.
16:07It works really well. Have you tried it?
16:10I I use Instinct every day religiously.
16:12Oh, Instinct's great too. I So my co-founder uses Instinct.
16:16Yeah, it's just it's all the same sort of similar things.
16:19Um yeah, actually maybe I should switch to Instinct. I pay for Grobot. It's like whatever. But anyways, I use Grockbot. Then we have like cursor. Then we all started using like factory for coding.
16:29Um I think the factory part that's really good is that you can like bring your own key. So we have a bunch of credits cuz we're like a startup from like Claude and Open AI and we just keep getting credits from every like VC and so you're just using like bring your own key for that. Um which has been very helpful and they're like $200 per month plan just basically unlimited tokens.
16:46Um Matan and Eno are amazing.
16:48Yeah. And I also think the app is better. I we used to be on Devon for a while um and Cognition and the problem with like Cognition was it was just so expensive like I don't know how but the tokens you use on factory end up going like 10x higher than they do on like Devon and it's kind of like this look it's they're like competitors and the same product but my like feedback to Devon is be like like why is it cost me so much to like use the product itself like factory is just so much cheaper um
17:17and I think also like the underlying harness of factory somehow is better than the Devon one.
17:23I'm thrilled to hear that. As a factory investor, I I completely agree with you.
17:27I just gave them an ad.
17:28Going to another one of your statements which you kind of touch on, but you said older GPUs aren't necessarily obsolete.
17:33What What do we not know about chip depreciation?
17:36Uh I think it's very simple that Core has also said like their A100s are still being rented out for like a 3-year period and the H100s are also being rented out for a very long time. People thought that they'd appreciate. If you asked this question last year, people would have said, "Oh, these ships are getting depreciated like tomorrow. It's like next day sort of thing." Um, but now people have realized like, "Oh, wait. The H100's are still renting out for 3 years and these ships are like 5 years old." Uh, the A100 is like I think actually H100 is 4 years old and the A100 is like 6 years old. Um, and
18:06they're still being rented out. So, it's like, oh, okay, maybe that's longer than we actually thought. I also believe that like the trip depreciation like the older chips will have a use besides being like frontier model training.
18:17They'll be used for lower scale inference, smaller model inference. And if you believe like inference as a market is going to split up between all these different categories like fast inference, slow inference, specialized, fast inference, slow inference, specialized, what do these mean?
18:32So I I think fast inference is that people want like their answers like really quickly. Um and so people just want highest tokens per second on their inference. There will be a company that specializes in fastest tokens per second. There will be a company that specializes in cheapest but slowest um tokens per second. And I think there will also be a company that specializ in like biological inference or like foul in a sense is like inference for like video or like image generation um specialized inference.
18:58That's so interesting. And so you don't think that it'll be a horizontal provider fireworks a modal that win?
19:05I think it'll be specialized.
19:07Yeah. I think it'll be special. I don't think one I mean maybe like maybe modal gets really good at like fast inference and like image generation inference um and they'll own those two verticals but I think it's like they like you own like one part of it what you said there with f like image and video what other mode requires specialized um I mean I think biological like people like in physics as well like these like the inputs are not just like text to text right it's like DNA or like proteins or like some some chemical
19:37formula to like some other chemical formula.
19:39Can I ask what market composition do you think the inference market takes over time? Will there be three or four big providers? Will it be an Uber, Lyft?
19:46Will it be an AWS, GCP?
19:48So, I think there will be like a horizontal provider that is like very good at like providing like base level inference. Um, that will be the everyday person's like inference company, but then there will be like a hundred different specialized inference companies that exist.
20:02There's like this joke on Twitter that there's like a thousand companies that do inference today.
20:07Yeah. Um, but I think I I believe like all of them will survive. They'll just specialize in one specific thing and no one's kind of found it. So, everyone's like, "Oh, we're doing like this type of infrance or this type of infrance."
20:17Do you not think it's a moment in time part of the cycle where there's a lot of new new people that come up and you can grow very quickly, but they'll consolidate as markets do and the actual number of players shrinks.
20:28I think the actual number will shrink, but at the same time, like it will be verticalized. Maybe I'm wrong about that. Maybe it doesn't verticalize, but like I I believe that it will verticalize.
20:38Can I ask you late cycle momentum investing in AI? For me as a venture investor, it has never been more binary.
20:47I tweeted this last night. You either get a round done in 2 hours or not at all. What are your thoughts on late cycle momentum investing in AI? I think it's a lot of hype if you can get a round done in 2 hours. like clearly like besides a lean investor, no one else is like actually understanding or looking at what they're doing.
21:05Um, and so I think it's a lot of like fearbased investing. A lot of it I believe and maybe this is not true about like all investors is like people didn't get into OpenAI or Enthropic early enough and now they're like okay it's a $1.5 trillion IPO or like two trillion whatever whatever the number they're like OpenAI is going to get to. Enthropic is like 1.5 trillion. Yeah. Um, and I missed out on that company.
21:29Uh, and so people are now like scared that, oh, I'm I'm not going to miss out on like the next OpenAI or the next Anthropic. And so everyone's kind of just putting their as much money as possible because no one wants to miss out on the next like large company. It's a fear of like missing out. FOMO is really what it just like FOMO investing.
21:46But it's not FOMO investing in the sense that VCs are like, oh, a tier one VC let it, everyone, all the other VCs try to go after it. is FOMO investing in that they see the public markets are like, "Oh, wow. Like, I need to start investing in this."
21:58Do you actually think VCs add value?
22:00I think I think some do, some don't, but some definitely do.
22:05Eloquently done. Um, what is overheated then? If you would look at a space where you're like, "Oh my god, it's like for me customer support." I'm like, "What the [ __ ] 15 providers."
22:16Yeah. Yeah. I mean, customer support's probably maybe one of them. Um, part of it's like the sort of specialized like things. A lot of people are like, "Oh, we're Rogo for this or we're Harvey for this." Um, I think those categories are definitely like taken at this point. I I've talked to Gabe a few times. I think Rogo is like killing it. Like Harvey's killing it. Lora and like those like the easy categories are all done. Like I don't think there's you're not going to compete against Rogo. You're not going to beat Harvey at like legal. If that is a category that exists, you're not going to beat them at it. Um, and maybe it is
22:46that open AI and anthropic beat them at it. But if you believe like these things will be specialized over time, there's there's no room for those. I think that's part of it.
22:54Okay. Where is underheated? You're a GP at 20 VC today. Well done.
22:59Where do you want to go and invest as a space?
23:03The as much of the energy and sort of like physical AI stack as possible. That's not even a crazy take. I think that's like a normal I take now.
23:11But let let's go deeper into that. What is that in particular?
23:15The data center space. Um that's probably because like I'm in the industry, but I've realized like over time that there are not enough um people that are building out products for the Neo clouds or the people building infrastructure. Like Cruso is such a good company because they build out all like the actual electricity infrastructure and they build out the transformers and like they virtually integrate very well. So if there's a company that can build like transformers tomorrow startup like I'm going to build better transformers, I would invest in that company.
23:43Do you see that often?
23:44No. Yeah. It's a kind of boring thing. That's why. Or or maybe it's like people aren't interested enough or they don't like like realize the market for it. But I think there there's companies that are very specialized in building out specific like products or tools that don't as a market doesn't look as big.
24:00Like if you look like the transformer market like oh like like how big can this be? But if you think data centers are $7 trillion, a lot of that has to go into like the physical building out blocks of this.
24:11Do you miss out by not being in the valley? Uh no, I think New York's like much better.
24:18We have this whole theory that people in the valley are like coin operated. Um in the sense that we say like it and this is true if you look at if you pay a researcher 100k more they're joining your company. If you pay them like two give them 50 bips more equity they're joining your company. In New York there's loyalty in people. there is like this like culture of like we are so loyal to like the people that we work with and the family and like there is much more of a commodery between like the people there compared to SF. Do you think that's true?
24:47Um do you know what I I think there's just a lot more group think in Silicon Valley than there is New York. It's so insular. It's so incestuous. It's like, "Oh my god, I'm going to FF and then I'm going to Bessemer." And I'm not saying I wouldn't take a tier 2 because I want the prize, but it's just like, "Fuck off."
25:05Um, yeah. No, I'm a real That was a great accent.
25:09I just can't I can't stand it. But I I'm going to be totally honest. I'm going to get in so much trouble for this. That's why I don't particularly love the American way of doing business, which is like everything is a transaction.
25:21And actually, that is not how I hate that.
25:24Yeah. I just want to do amazing work with people I respect and sometimes there's no actual money to be had.
25:30Yeah. Yeah. My I think part of like when we started the company it's part of the reason it's like oh this like market or all these things but I get to work with like my four best friends from school every single day. We like all went to college together. We're all in like fraternities together. It's like there's no better life that I would like live besides like working with these people.
25:52Um, and that's like just like a very like common reason like I think people start companies for all different reasons, but there's no other place in the world where I get to work with my four best friends.
26:01I totally agree with you. I think that's really romantic and sweet.
26:05Um, so why do on what's the like I have to do this because uh part of it is like I I really wanted to like I I think throughout college if you ask like my friends or like investors or other people they'll say like oh he was like definitely going to start a company in that like I did all these jobs and like it was like whatever but it was really like I spent a summer at BCG um and I was on like the private equity team and I actually learned a ton throughout it but I was using I was like coding um and chatt had just come out and so what I would do was like I'd
26:34figured out like Chat GBT is very good at like making Excel formulas and all these things. So you could literally like the MD would give me a task like oh go like organize these like filings great I'm going to take chatbt give it to that it would come out with the Excel formula you paste it into Excel done I would do it before like the VP or like any of like the senior people could do it and people were like wow this guy is like good but it was not that I was good it was just that like I was just I was just so good at using like chatbt to figure out like the coding and the
27:02formulas and then I was like writing like Python code um cuz I studied like I was I was a CS major in Um, so I could like code and I was like writing like Python code for them and they were like, "Oh, wow." Like I've never seen someone that can do this cuz they're all consultants. And at the time I was like, "Yeah, I just can't do this.
27:18It's like a full-time job." So I like did a bunch of things. Uh, it must have been really sad to lose you.
27:23Yeah. Yeah. And I worked it was I actually told them no. And it was funny. A recruiter like the recruiter called me in like March of this year, like April of this year after A16Z did our round.
27:32Um, I was going home for like a few days and they called me and they're like, "Hey, like why aren't you not at the office?" It was a Monday. I was like, "What do you mean? Like I I quit like two years ago. I was never coming back." And they're like, "Oh, like apparently it says you're coming back and all these things." And I was like, "Yeah, sorry.
27:46Like I I'm doing like I have a whole different startup now." Um, it was a very funny experience. Yeah, it was a That that that is hilarious. So, if we get into it, why does compute actually need a market? It is purely just supply and demand. That's how prices work for compute in the sense that there are many suppliers that sell basically the same good which is an Nvidia GPU and there are many demanders who are demanding all the same good with different parameters associated with it. Some people need the
28:16location to be this. Some people need like 3 terabytes of RAM that are two terabytes of RAM but it's all kind of in this like same sort of bucket. And if that's true that means there needs to be a market for this because there's many suppliers supplying the same thing. there's many demanders demanding the same thing and that's just purely a market.
28:32If it's simply supply and demand, is there no moat in this market?
28:37I don't think NeoClouds essentially have a moat and people are trying to trying to find a moat today. Why do you think like coreweave is like, oh, we're doing post training? Oh, we're doing like tokens.
28:48Well, they have to expand product to gain mode.
28:50Yes, that that that is what everyone's doing. They realized like early on there is no mo. Um, and so now everyone's trying to expand to find some moat in Is that paradoxical to what you say earlier though, which is like I would buy them, I would be long, and then they have no moat.
29:04Yeah. But that's I think that's fine. I think there's still like they're all good businesses. Um, I want to be like the landlord. Like that's a great business to be in. Um, but yeah, not I mean I don't think not having a moat as a like the moat for the neoclouds is like I own the infrastructure. I I I completely understand you. We have demand, we have supply. Why have we not had a market for them before?
29:25Uh partly because the suppliers there weren't that many. I think early on in CPU land there's only three suppliers. AWS, GCP, Azure, Oracle maybe. Now there's 300 NeoClouds.
29:36Okay. And so we have massively fragmented supply side and that means that we don't have a market.
29:41No, that the fragmentation means that there is a market for it because there's so many different suppliers for it. I also think on the demand side now there's 100x more demanders than there were for CPU products. Um, and they're all kind of asking for the same thing, all trying to go after the same gold vision, like, oh, I'm going to build this, I'm going to build that. Um, and so that creates literally like a way for people to bid on the price. GPUs today, every every all the neo clouds have now started, what do they do now? Auctioning u for the chips, which is like it's a
30:09market. What has to happen then if we think about it compared to a traditional oil market? What has to happen then for compute to trade like oil does? I think it today it will trade a lot of people like so like we we build index products for compute itself and then we do that by running like a physical market where we use the data to actually build the index products and similar to oil where if you look at Brent and the Nordics people always like see how much of Brent is actually like of the global oil volume is Brent and it turns out it's
30:37like I think it's like 08% 7% of the actual global like oil volume which is such a small fraction of the total oil that exists in the world and if you think of compute it's very similar to Because compute there's all these different flavors, these different gradings of quote unquote like crude oil of a B300. But what you really need to figure out is you need to track a very specific type of that compute and that becomes the index or that becomes like the default of which you track. It doesn't have to be all of them but it has to be a very specific type and then
31:06that becomes a reference price that people everyone sort of mentions and that's becomes the sort of standard the gold standard of what compute is. And I agree like I think Eric like said probably like last week um or a few days ago he's like oh computers like oh there's all these different flavors. I totally agree with them. There's all these different flavors. It doesn't mean that we have to somehow figure out a way to standardize all of them. Saudi oil is not the same as Venezuelan oil. Um and they all trade still like everyone still quotes off of like the brunt price or the WTI price or whatever that may be.
31:36So what is it then that is required to make it trade? It's like a standardization. It's it's picking a very specific type of the compute and working like just getting that compute standardized.
31:47Huh. How much of it is trust and relationships?
31:51I think a lot for the trading part of it, it's fully trust and relationships.
31:54Why? Well, if it if it's like as numerical as demand and supply and auction, it is. So, it's partly that, but what is the mode that S&P has or what is the mode that like a Moody's has? It is like trust and long-standing business.
32:08I totally get you. What is different about your business compared to an S&P and Moody's building in a different market? We're earlier and we're faster moving and hungrier. Um, that's the that's what we say. But I think reality is like we we say like we want to be at the S&P for compute. That is like the part of like the like that's the scale we want to get to. Even larger scale than that. And part of that's because I think compute is a larger market than oil will be or any of the other commodities have been for the past like 100 years, 150 years.
32:37And we really want to build the business in the same way where they run the physical exchange and they use that data to build out the actual index and then you build a bunch of data products on top of that and then you do ratings and all these other things on top of it.
32:48Well, pause pause that you're expanding product suite there. How do you know what to start with? Why do you choose that and why do you choose what's coming?
32:57Yeah. Um part of it we started with the index product because we think compute needs to be traded on a public market. There needs to be futures, options, all these sorts of financial derivatives that sit on top of a GPU.
33:08Okay. So, it's indexed on top of GPUs.
33:11Um, and in order to get that data, you need to run like a physical exchange, right? A spot, like you need to run Brett in the Nordics for compute. So, we built that out. Um, and that requires like figuring out, oh, how am I going to get chips? So we obviously don't have a ton of cash and instead of building a neocloud it's like how do I get other people to give me cash to like they can they invest in chips and then I monetize their chips for them sort of running like an Airbnb style model. Um, and so we started doing that where like we find like other people to do it and we'll do
33:41like a sale lease back or all these sort of different like financial ways where like and then a day like we will control the chips but someone else like ends up um investing in the chips or investing in the data center project and all these different ways to do it.
33:54What did you not expect you you have come up against and found challenging?
33:58Uh going like this deep. I never thought we would have to like go all the way to the chip level. I thought we were just going to be like a software data company and now we've got it.
34:07Why do you have to go to the chip?
34:08There's no one that's gonna stand give you their data or standardize it for you. So in order to get the data for the index, you have to either go to like a bunch of NeoClouds and say, "Hey, please give me your like transaction data of what you sold this for. You think Cory is going to give you that? Hell no."
34:21Like they're going to be like, "This is this is the most like valuable thing in the world for me. I'm not going to hand this over to you." We tried for like six months like, "Oh, can someone please like give me their data?" Uh and we had some like to be honest like we had some people actually give us their data but at the end of the day we want to build a very robust market and part of that's like we need to have billions of dollars of data and having billions of dollars of compute is you have to either go to a neocloud or go do it yourself and we have this concept of like no one's going to come save us. We got to save
34:50ourselves and part of that we'd have to go do this ourselves and I didn't realize like you asked me a year ago we'd have to do this. I didn't realize we'd have to but now we kind of do. Does that inhibit the speed of scaling?
35:01No, because you can just get other people to put their money for in like I think there's a ton of people that want to invest in data center projects. They just don't know how to monetize it. They don't know how to do all these things.
35:11And so you kind of just teach them like they're like, "Okay, if you put this much money here, this is what the IR looks like. This is where you can return it. I will control the trips. I will manage it. I'll run it for you. You just put the money here." Um they're like, "Great." It's kind of the model. Yeah, it sounds very simple. Obviously there's so much complexities to it but that is like the simplest way to explain it.
35:30No, I I totally get that. Who will be the first mega traders in compute futures?
35:35A lot of it I think it is the people that are exposed to the price of compute. Uh and the way they're exposed is they're exposed to like the credit risk of the people buying the compute.
35:44Right? If I lend out my compute to a startup that has just raised a Neil raised like $100 million and the compute they're buying is really like $250 million, my bet as a lender is that okay, this team is going to keep raising venture dollars hopefully and that venture dollars will be spent on this today. Many of the Neolabs don't make money. So, it's not like they're going to like generate like revenue or profit to hold this or my bet is that the Neocloud somehow um the price per compute goes up and now they bought it
36:13for very cheap and they can sell it for higher and then they're going to pay me back. So, you're really exposed to like the company itself that you're lending to. And if they default that means I have to go in the market and now sell that compute for a certain amount of price which could be lower which could be higher than what it was before. So, that's like reason one. Reason two is that no one knows what the cost per GPU hour is on the depreciation cycle. And a depreciation cycle is like, okay, 3 years from now, what is my cost per GPU hour going to be? I need to basically
36:42get to four years to pay back my loan or I need to get to 5 years, but I only have a three-year contract. So, you're kind of exposed at the end of it. And maybe your contract is all the way to 3 years and it covers everything and you get the lenders and you have the chips.
36:53I still need to make my money 3 years to like the end of life and I can lock down my money now. The same way farmers used to do this like back in the day. If you look like a 100 years ago, like what did farmers do? They needed to plant corn or whatever that product was like they needed to sell futures back in the market because I would plant today and I only realize my revenue a while back after this and I don't know what the price is and if I make a loss I've just wasted like six months of time and money and all these things.
37:20You're only as good as your underlying tenants as you kind of said there. Um, with that in mind, is Meta's bond too expensive? The health of their business is phenomenal. Anthropics bond price, you know, which I think Blackstone is kind of the regulator or the provider of is is actually not amazing. These are both fundamentally good business. Are they super expensive?
37:41I think Meta is like getting I mean Core We is way higher than Metas is um to be fair, but it doesn't have such an underlying good business.
37:50Yeah. Yeah. Yeah. Metas is partly like I think it's getting more expensive just because like people are just like a little scared in the market. Um and also the amount of like growth that people money or people are putting into the infrastructure. If you look at all the hyperscalers, they're lally taking every single dollar they make on their regular business lines and just putting it straight into data center capacity. And so people are like wow like I've never seen like this much of like money being put into a project before. and a lot of public market investors and and you're moving from positive free cash flow to negative free cash and
38:19Google and Facebook which is alarming and it's partly like people in I mean I live in New York but people in like New York are don't realize that how much AI like what what really is this thing like is there really demand for this and there people are just getting like slightly like concerned about that can I ask you and then this was a brilliant one I think from Mars actually but he said when you look at the index today does an index drive like standardization of compute clusters and how do you think about that?
38:46I think the index does drive standardization of the compute clusters and it's a good thing because the way standard I mean Nvidia Jensen themselves will say Nvidia clusters are funible.
38:55They what they mean by funible is that if one tenant backs out someone else will take the same cluster. It's not that you're doing a very specialized build for every single person. The index itself will drive standardization which I think is a good thing for the credit markets or is good thing for any of the lenders because if the product is standardized that means like if one customer defaults someone else can move in. If I have an apartment that all looks the same for every single building that means it doesn't matter if like I lose it someone else can just move in because you're just going to get the same thing.
39:23Yeah. And I guess it would remove the discount rate.
39:28Okay. I I didn't see that before. So given how much time you spend with people who buy and sell compute on a daily, what do you see that you don't think other people see enough?
39:39The prices are people are willing to pay um now just like gone up way higher. Um are the buyers changing?
39:46The buyers are definitely changing. It's going way more to as to your point earlier a modal fireworks a foul. Um it is going to people that actually have the cash like FA just raised like 1.6 billion. Modal also raised um what is it 800 billion or 800 million something like that one?
40:04Uh I think they raised 750. Yeah.
40:06Yeah 750. Um all that money is just going to compute and that's way more than a small like startup has. So part of it for us our job is like making sure that the small startup still has access to compute while feeding these like larger customers.
40:20I totally get you. Are you able to do both at the same time? Does it make sense to do both at the same time if modal can just pay you 100 million bucks? It does because without that there is no market, right? You have to somehow seed like liquidity in the market. If modal is the only one that's buys, all the competitor is going to consolidate. No one else gets access.
40:35What's the hardest thing about seeding liquidity?
40:38Um, you have to say no to a modal. You have to say no to a fireworks. You have to say no to a foul which is willing to pay you gazillions of dollars in revenue and profit for your underlying business to like kind of look in the long term.
40:50It's like short-term game versus like long-term profitability. How much is a market ultimately worth? Like if you think about like ICE today and when you look at like or in five years time being the kind of market for compute, how much is that actually worth?
41:05So I I think partly it's like we want to be like we're a data index company and I think the market is part of it. Um and like look a lot of the team today focuses on the market of it because that's how we actually build the data and index product. It's very like it's funny we if you look online all these things like you're like oh this is like a data and index company and that's true and that's like five people on the team six people on the team um but in order to get that data that is like the other like 30 people on the team to actually make this like team like look really
41:33good uh I mean ICE is a hundred billion dollar business is like $120 billion business uh S&P is worth more than both of those I think they're like 150 $130 million business Um, and so these are all like hundred billion dollar businesses plus. And they built their businesses off of like electricity and oil and um, traditional commodities. But if you think compute is like going to be the next largest commodity in the world, it's going to be much larger than all of
42:02these. So the market opportunity could be in like trillions.
42:05Especially as we move forward to like an inference market. When does inference not training set the forward price to you?
42:12Uh, I think it does today. I I I would argue many of our customers today use a lot of their compute for inference um itself that they're just like buying compute to serve like tokens off of it.
42:25But the good thing about the inference demand is it's not speculative. It is like real. The tokens have to be used by someone. Maybe it's a startup that uses them. Maybe it's like an enterprise that uses them. But it's it's real. Like someone is like paying on the other side. Training is much more speculative, right? It is like kind of like biotech where you spend all this money developing this drug and you don't actually know if it works out or doesn't work out.
42:46So if you think about like mispriced risk in the market today, where's the mispriced risk in the compute market?
42:52I would argue that it is a lot of these um Neoclouds that are not actually looking at the tenants that they're lending to and just like saying, "Okay, great. This company's raised like 100 million have the compute." um and they don't actually like do the enough diligence and before lending out the compute and inevitably the way the world works is many of these neolabs many of these companies will end up failing startup like success rate is very low and today all kind of all boats rise with the tide is like the saying and
43:21that's kind of true today but at some point I think a lot of these companies will not succeed and these neoclouds will be left exposed without a way to hedge I'm really sorry given the demand access yes will they not just fill it instantly they will. But the part of it is like which model is going to like win, right?
43:38There's a lot of Neolabs that are all going after the same thing. Um then you kind of like have to believe like some will do well, some won't do well.
43:46What do you think will determine the winners in Neoabs versus not talent? Um I think everyone has the same access to the tools. Um also I think it's talent and compute. If you have more compute that means the models will get better. I mean it's a very known thing. more compute, better models. Better talent means also better models. So, it's a combination of both things.
44:08And you can maybe even argue that having more compute is better than having better talent in the company itself.
44:15It's such a big mo more if you have the compute, you just kind of like Why do you think like SpaceX berser?
44:22If you could have more compute or better talent, what would you choose?
44:27I think the the models are good enough to do like RSI. So you just kind of just need compute or they will be good enough in like 3 months from now.
44:35You know, I think everyone is so challenged and the markets are Shopify dropped 15% after a tiny miss in earnings. Like tiny miss. Um and everyone is just like unsure of the 5year. Like for the first time in a long time, everyone's like, I'm not sure where we're going to be in 5 years across so many different markets. And that makes pricing anything just incredibly [ __ ] difficult.
44:56It just increases volatility. Yeah, that's what it is. So the options are honestly like you could argue they're underpriced.
45:01And my question to you is do you believe that things will be as significantly changed as people think they could be or do you actually think that things take longer to adjust?
45:13I think the enterprises will take longer than people think to adjust and then they the world will change faster underneath them. I think like the San Francisco bubble will change way faster than like the rest of the world will like the rate of progress. the San Francisco bubble is like the AI world is what I call it that. Uh and is that the the model the capability of the models would be so much better and so much like it'll be exponentially better while the world is still like stuck in like a very linear growth curve because the enterprise is like oh like I have this technology like how do I put it in my business like Coca-Cola is probably
45:42wondering like oh how do I put like AI in my business like I have to use this somehow and people haven't really figured that out. What do you believe today that everyone thinks is pretty [ __ ] kooky that will be very common in five years?
45:55I think like humans will be like a biological vessel for like AI. I think the AI will be like so much smarter than we are that the role of like humans is to like basically um not even control the AI but like teach it like biological like feelings and emotions and all these like things besides like being the sort of like parent of the AI. I think also longevity stuff. Um I really believe like I think
46:22dying will be a thing that maybe isn't like real in like a few years from now.
46:28I believe like the drugs and the technology of like biotechnology will be so much better than it is today which ends up meaning like the average lifespan for humans will go from whatever it is like 80 something years today all the way to like 150 200 years and the quality of life will also increase. It's not that when you get to like 80 you can't do anything. You can't like stand up and when you're like 100 you're just like sleeping all day. I don't think it'll be that. It'll be like you're living the same life I'm living at like 23 today that you will when
46:57you're like 80 years old.
46:58Well, Brian Johnson would be very proud.
47:00Um, can I ask you what revenue line do you not have today that you think will be very prominent in 5 years time? Um I wouldn't say that we not have but I think like uh debt ratings for like data centers um will be large for us going forward.
47:17Sorry this is where you are the like certified provider of a debt rating for Moody's for data center debt and the inherent value of that is for all lenders then you become this trusted Yeah. And part of that's do you think Moody's understands how data center like debt works?
47:36Who knows? Like, do they really understand like the credit risk that Modal is having on their people that they're buying the infrastructure from?
47:43I don't know. Um, and it's funny that I say it's like Moody's for data center debt ratings cuz I just think it's just like data center debt ratings. Um, and Moody's has made such a good name for themselves in like ratings in general and so is S&P and many of the other businesses, but I think for us like we want to get into ratings soon. Not today, but soon. Do you find it difficult to retain focus?
48:06Yes, there's so many things to work on which I think a lot of companies just like just keep hiring more people and just go into all these different verticals.
48:14If you were unconstrained by resource, what would you do that you're not doing?
48:20Everything. Um I just try to do as much as possible. Like honestly like we would start building out like literally physical data centers. Like we would just go like full like tell like Cruso we're competing with you. Well, you're seeing the inference providers do that. Your fireworks of the world do that. Do you think that will be a successful play as they eat more and more of the vertical chain?
48:41Um, I think they will, but they're going to be way slower to do it than the other NeoClouds are. Like the Neoclouds have like gone through this cycle like I think two cycles now. Deploying chips, learning things, deploying chips, learning things. Um, as a company, we're like basically like halfway through a cycle of this. I think the more cycles you get through the more you learn and you figure out things. I think they will be successful, but it's saying that okay, are you going to be able to beat coreweave um in their own inference
49:09game? Like are they going to be able to get to inference faster than are you going to able to chips? And I think I would take the side of like a core or an EBS and say they're going to be able to get to the uh inference before you are going to be able to get to the chips. If I had to pick I I said to you about if you're resource unconstrained, you kind of are resource unconstrained in cache.
49:32Um I mean you've raised when you you didn't need to. Um why raise?
49:39Uh part of it is getting the right partners around the table. We So the way we build the business is like we want to build this as like a family business.
49:46Like we believe it is like truly like a um like we are a group of friends building a business together that will like last for like generations is the sort of like idea. It is not the um I think there's different ideas to build a business. It's not like I'm going to build a business for like 5 years then I'm going to go sell it and make a ton of money. Like this is supposed to be like a a long-term play for us and we want to build it like a very long-standing business.
50:11Uh I used to play like League of Legends and it's just a character and I think it was a very cheap four-letter domain. Okay. But yeah, that's the honest answer.
50:21No, dude. That's fantastic. I I I you know, sometimes you think like it's that Greek mythology and this kind of very thoughtful. Well, League of Legends works better in this game.
50:30Yeah. And and and we literally So, what I did is took the I realized like Andrew and Palanteer all from like Lord of the Rings. So, I went through the list of like all characters from like uh or like people from all the sort of like cool like games. I was a huge video gamer.
50:44So, I went through all the the game things and League of Legends you can go through all like the the players and I found out what was the cheapest four-letter domain. And there's a study that starts with like shorter names are five times more likely to succeed. Um, and so I was like, "Okay, so we need to go like five letters or below." And so I just went through like all of them and I found that, "Oh, this works." And so it was just like a a cheap four-letter domain.
51:07You you chose Miles, one of my favorites. Can you just take me to like why did you choose Miles?
51:15When we met Miles, the first like 45 minutes, he literally didn't care about the business. He just talked about like us as people. And so I know when he invested he like cares about the business and the idea and all these things but he cares more about us as like people.
51:30How do you That's so interesting because I also spoke to I think it was the Vine guys. Um I think it was Alex who said that's wonderful.
51:38Yeah. who said that he went for dinner with you with Adam Falcon and Adam was trying to talk about tennis with you cuz Adam plays tennis and you play tennis and you were kind of honestly quite uninterested and you were like glazed over and then you were like okay kind of enough of the small tour let's get to business and and so I'm just like that's intriguing because that's seemingly different. Yeah, I I think so this is you just not feeling it.
52:06No, no. So you you So the story is like uh my co-founder Wayne is like a pro tennis player and Adam also played and Wayne played at MIT, Adam played at Harvard and so they started like talking about it and Alex and I are just like sitting there and there this is like 10-minute conversation. um we're like we don't even know to like jump in and at the time I had another VC meeting right afterwards and I like I genuinely like I told the way we met Vine was Alex LinkedIn DM' me want to like chat I called him I literally didn't take a Zoom call cuz I was like this is
52:35probably not worth my time so I called him on the phone I was like oh this guy's like pretty good like maybe I should like talk to he's like oh you want to go for dinner at this place and I was like oh wonderful it was like a one minute walk from my apartment and so I had like gone and I had like another big VC meeting I think like large fund um that also gave us like a term sheet and I was like Wayne and I were just like negotiating the term sheet at the time. We're like we'll probably take it and go knock out with these guys. And so I was like all right let's just get to business like see what these guys offer us. It turns out the bond offered us like a term sheet. It was like same
53:05price. Um and I was talking to Wayne and I was like okay like how do we want to do this? Like if we go to this like large fund we were just going to be 01% of their fund 0001% some small fraction. are they really going to care about us? Um or else we go to Vine, it's like we're going to be like 5% of the fund or like three whatever it may be like some significant amount. Um and at the time we like realized I was like okay like Adam and Alex like really care about like us as like people and they're
53:34like way more interested in like who we are rather than the idea and for them this will matter like way more than it does for like other people.
53:42That's so interesting. But I because it's different to every other young founder that I meet who just is magnetically attracted to the brand.
53:50Yes. I think that it is a short it's like short-term gain versus long-term gain. And you just have to be like disciplined enough to like choose the long-term game.
53:58We're sitting there and we're both founders. Dude, I've got I've got a term sheet from a big like multi-stage uh and I've got a a boutique. Would you advise me?
54:07I'd go for like the boutique and that's because you guys care more. I do all these like reference calls for these investors. The boutique investors like one care about you a lot more because their success is tied inherently to your success. I've also realized like a lot of the partners at the boutique firms are there for more of like the mission.
54:23They're they're also in like a startup, right? A boutique firm is kind of like a startup. Like they're also there to like I need to make a ma name for my like VC firm and part of that's like I got to work extremely hard to make sure these startups succeed. So I would I'd really go with like the boutique firm.
54:36What would be your It's very contrary, right? I think to the world today. Um if like it's hard to resist but I I think it is the right decision to make for everybody.
54:45And then you raised from Andrew.
54:47Yes. And then we did the big like firm.
54:49How did how did that go?
54:51Honestly like we we were like talking to another like large like multi-stage firm at the time and they were just like taking very long time. And so uh it wasn't even me. It was like Wayne.
55:01Wayne's like all right we just got to like ask like we got to like start like moving this. And so we went to like um it was like Chris Dixon's team and talked to them for like a week, got the term sheet and we ended up going with them. God, I bet the other firm were like, "Oh, no, no, wait, wait. We've got one. We've got one."
55:16Uh afterwards, yeah, it's brilliant way it works. Um and then thrive. You chose Thrive over others.
55:24That's because of Miles or the Thrive brand?
55:28I I also think like I mean look, the Thrive brand is great. There's other brands that are also like great, but I think the partner matters like way more.
55:35Like we are a people company. Um, and I think it was Miles and we also met like Nithan on their team and I had like lunch with Nathan. It was like the most like it was the best lunch I think I've ever had. Like I learned like so much for for me as like a founder. It is so much about like I'm so young that I'm going to make a ton of mistakes and I know that like if you talk to like other founders that are older than me, like if you talk to like Eric at Modal, he he's already figured out that if I do this I'm going to fail. Like I don't know that yet. Like I I don't know all these like I was never in industry really
56:04before this. So I have like so much more room for mistakes. I need to minimize that. And part of that's like just talk to as many older people as possible and like learn as much as possible that you can't find in books or texts or all these things that you learn these like people skills.
56:18What do you know now being the short while in that you are that you wish you had known at the start about being a founder?
56:26um the mental like stress and like the level it is uh there's like a saying that it's like it's very difficult. I think it is like very difficult and it is like mentally consuming and like taxing because every day it's like this you're just like in this like struggle like oh what's happening next? What's going to go on? What's going to go on?
56:44Um I don't think people realize it when like getting into it how difficult it's going to be. But that is why we got into it. We were so naive that we thought this was like so easy. We're like why is no one doing this? this is going to be like the easiest thing we've ever done.
56:56Um, and then you go in and you're like, "Okay, this is why no one else has done this before because of how hard it is."
57:02But I think it's like the level of like grit it takes is the the volatility that you will feel in 24 hours of I'm the best in the world and this is going to be amazing and then this is not going to work and we're Scrooge is is like in the same day.
57:16Yes. Yes. That that is what you have to get used to. You have to get used to like emotional volatility which is so hard for people. Um I people there's like this thing like oh like what do like startup founders need? I honestly think like if everyone could have like a therapist it would be like the best thing.
57:29Do you have a therapist?
57:32But but I honestly I wish I do. I'll if our investors are watching someone go find me one please.
57:36Oh dude. Dude I'll find you a therapist.
57:38That'll be my gift to you. She'll be like Wendy Rhodess from Billions. Have you seen Billions?
57:44Oh dude come on. You seriously? Oh it's a brilliant show. Do you watch TV? No.
57:49Uh not really. No. I'm going here.
57:53You don't date. No. So that so that's maybe the So there's there's this like huge joke that um the reason I say I don't date. So um the vine guys got us these hats that say single till series B. Um and series B is like next round.
58:07So I'm like okay this is like soon like we can't be single after series B. So I'm getting him new ones that say single till IPO. And so my joke is until I return like investor funds I'm not going to date. It's not it's not not actually the reason why, but we'll we'll go with that.
58:21Can I ask do you think that dating does actually just distract?
58:25It definitely does distract you. Like you lose like the grit um like the the the anger inside of you or like the like I got to go do this, I got to go do this. Um but at the same time, it also helps like emotionally regulate you, right? Um humans are like emotional people and so part of it is like being able to regulate like all the the volatility with like stability and like decreasing the volatility over time. Do you care about money?
58:49Uh, I don't think so. I I think when we started the company, I used to a lot. Now at this point, it's like no. If it was like, oh, you make no money in the startup, I think I would still do the startup. It's I think it's the people that I work with that I there's just like so much fun.
59:03So, what what's it about for you? So, like I can be working with my friends.
59:10That that's literally it. There's no like there's this like concept of like okay if you believe that when you're living your whole life as long as like I'm comfortable and like I have I'm at the age right now where I have no one else to take care of right it's just me u and I live in like a sixth floor walk up and like there's no like what is a sixth floor walk up so in New York the legal amount of like stair there's no elevated in building so the legal amount of stairs you can get is six floors before like it's illegal in New York to have like make people walk up
59:38you have six floors of stairs Yeah. Six floors walk up. Yeah. Me and one of the people we started to come to, Drew, live like a sixth floor walk up.
59:48Uh I mean it was just like an apartment we found that's close to the office and I was just like why not?
59:53It's like fine. Yeah. It keeps us gritty. That's what it is. Keeps us hungry.
59:57You're not going to have burglars.
59:58No, we're not going burglars. No. No.
1:00:00It's It keeps us hungry for warmth.
1:00:04That's That's bizarre. You live with someone?
1:00:07Yeah, we have two bed.
1:00:08Dude. Wow. I think a lot of people think like, oh, like your company's worth this much, like you must have like so much money. Are you missing? Like, no, not really. I never worked a job before, too. Like, I don't have like savings.
1:00:20I I completely understand that. Now, would I if I was you and your co-founders take a million bucks in secondaries when there are so many people who want to give you money?
1:00:31Yeah, I I I would take a million bucks off and I Maybe we should. And and and again, I'm not advocating for private jets or [ __ ] but like a great gym, healthy food, and a nice We do all that. We do all that. We have like a gym membership and we all like we're all like But like a nice apartment. Yeah.
1:00:49Yeah. But like it does it really change things? Like I I spend like most of the day at the office on the weekends like I I'll go for like a run on the mornings and I'll just go back to like the office. Like there's like nothing else to really like do.
1:01:02Do you think you have a good culture?
1:01:04Yes, I think our culture is by far the best out of like any startup. So, in the company when we started, I really didn't think work culture was important at all. I thought it was like you find the right people that are smart and you work. I've realized now over time we made mistakes.
1:01:17We've done all these things. It's like work culture is so important because you end up spending 60 hours, 70 hours, 80 hours a week with these people.
1:01:25Sometimes with our like engineers sometimes, uh there's a guy named Will on our team. we spend like a 100 hours a week together. Just me and him like figuring out issues and going through like so many things together. Um you're going to spend this much time with this much people. Like you're spending more time with them than you're spending with your friends, your family, anyone else in your life. You spend more with the people you work with.
1:01:44That means that the people we work with has to be so enjoyable that you never want to like do anything else. So, we like care so deeply about like culture of the team that we've hired people and the people that we've hired are all like ex- college athletes um or have played some sort of sport growing up. I heard in like this like I don't know who said this but someone someone said this was like when you grow a team from like 20 to 50 to 100 people what gets lost is like communication between the team like you get on these like middle management
1:02:12layers and all these things and to solve that problem to keep like if I go from today we have like 28 people 29 people if I want to go from that many people to 100 people in the company I need to figure out how to for like 3x my team while 3xing like my revenue numbers or 3xing like the productivity of the team to do that. I need to find people that are really good at communicating with each other. It turns out the best like communicators are like college athletes because what do you do when you play a sport? You're just like communicating
1:02:40all the sing every single time. And the second thing is like you are so competitive that you hate like losing.
1:02:46Like you want to find people that are like competitive enough that if you tell them, "Hey, competitors launched this feature. Why don't we have it?" They're like, "It's we have it tomorrow." Um that type of like level where it's like I just want to win. Do you know why I like athletes? I like athletes because you have to consistently do an activity that has no immediate benefit. And this is specifically like gym. I like people who really are very fit because you have to consistently do an activity that has no immediate benefits. You don't
1:03:14suddenly get jacked after five sessions often on your own with no inherent motivation for a delayed payoff that may come later.
1:03:23It's like it's like short-term game versus longterm game. Like you got to take the long term. And and when you're going to the gym to get jacked, you're basically either doing it cuz you're insecure in yourself and you want to look better or you desperately seek someone else's approval. It's one or the other. I don't give a [ __ ] but I like both.
1:03:39Or or you just want to be a healthy person and live a long life because you're not getting jacked.
1:03:44Sure. Sure. Yeah. You're running a lot.
1:03:47Yeah. You're running a lot. Whatever.
1:03:48You're getting the Brian Johnson workout.
1:03:50Yeah. Yeah. Exactly. Um that's so funny.
1:03:52Anything else on culture and team building that you've learned that you do that's weird or different?
1:03:58Um the interviews we like take people dinners and like weird uncomfortable situations to see how they like interact or act.
1:04:05What's an uncomfortable situation?
1:04:06Like we'll we'll find like people on the team that have no like we'll take an engineer with like go to market people and like have a dinner with them and then we'll like we'll just see like do they vibe with the team. I think we have like two rounds of interviews where it's like technical and we'll see like can they do these like things but with like AI tools I think a lot of people are like technical enough and we're not doing frontier AI research this is like it's technical in the sense that it's technical and like a very specific domain but really it's you got to find people that are willing to be mobile and can like navigate well so can they like
1:04:34handle uncertainty bit of a hard one you you're there's Axios reports about you there I see you on Twitter a lot yesterday specifically Um, and I'm not actually going to the material, but I'm sorry if I'm this it sounds patronizing, but you are young and the numbers are big. I've seen this before where people believe the hype. How do you kind of check yourself? Revenue numbers, profitability, look under the books,
1:05:03tell me if my business is more profitable than any of like the companies that exist today. And going back to like what I said before is like we are a long-term family business. And part of that is like our numbers have to like check out with what the valuation of the company is. And today I think our numbers do check out. Our multiples are much lower than any of the other companies that exist today. And that's because we've gone through this like process of fundraising that is very like disciplined and like structured. We went with Miles even though we had term sheets that are way higher than that. We
1:05:33went with like our investors that um gave us like term sheets that people people gave us term sheets of like two billion or one and a half billion and like crazy like numbers and we just said like no I think we're like going to be disciplined at this point and just go with that.
1:05:47Also it says a lot about the investor if they're willing to like slingshot a $2 billion term sheet.
1:05:52Yeah, I I think so. Part of it.
1:05:54Okay, we're going to do a quick fire. Um what was the single best VC meeting you ever had?
1:05:59Great. Um, can I can I exclude like our current investor list? All right. Well, we we'll we'll exclude the current investor list.
1:06:07Um, so they don't all get mad that I don't pick one of them. Um, I met Nick CR at AIX when I was 17 years old. I ended up investing in the last round or this round that we just did. But the reason that um I've known Nick for a very long time is he's mentored me since I was 17 years old at MIT. I was a freshman when I met him at a coffee shop. It's called Flower Bakery on like um Nash Avenue right next to my team and I met him at a coffee shop. He gave me advice for like 20 minutes. He used to work at Great Croth at the time.
1:06:35He's moved two different firms at the time and I've talked to him like probably when I was still in school every like few months. Um now I call him like once a day. It's crazy. I And the really thing I enjoy is that he cares about me as a person rather than like the company. That that's like the most important part.
1:06:51I love that. And AX is Rich Socher's father. Yeah. Yeah. I like him. He's good. Um, you can add one person to your board who's normal.
1:07:01Wow, you really thought about that one, didn't you?
1:07:05Uh, I have not. No, I really want to meet Travis. We got to make that happen. So, why don't you Well, I I haven't asked anyone. Maybe I should. I'm like a fanboy, so it'll be really awkward if I meet him.
1:07:14I I think I think with a clip after this show, you might be able to make it happen, my friend. Um, what would you ask him?
1:07:20Um, one question. So, how do you handle emotional volatility?
1:07:25That's a that's a good one.
1:07:27That's a really good one. Okay. You've got Anthropic at 1.5. You've got Open AI at let's say 1.2. Which one would you rather buy?
1:07:35Open AAI. They have more compute.
1:07:37It's that simple. Yeah.
1:07:39But anthropic have the cash. They can just buy it.
1:07:42You can't I don't think you can spend C.
1:07:43Like, how much are you going to spend on it?
1:07:45Wow. So, Anthropic will be bottlenecked because of that.
1:07:48What does that mean? I don't think I mean you can spend as much money as you want. You can't make the fabrication line at TSMC move faster.
1:07:56So what happens then they are not able to provide the products they want to.
1:08:00I didn't think yeah it's sad and like the models will be slower. They'll be there be like less like growth than open AI will. I think look the models for open and thropic will one will I think what is it tragedy soul six better than uh fable 5.1. I I don't know what the benchmarks are, but maybe Fable 5.1 is better than uh six or Astra is better than Fable 5.1. Um but that changes every like two weeks when new model comes out. Where is Jev in 5 years time?
1:08:27I think I Okay, so I haven't used Jev.
1:08:30Um maybe I should. Diego might be mad when he sees this. I haven't used Maybe I'll go home and use it. Um so I don't know the right answers, but I've got him coming on the show in a couple of weeks and I need to sound thoughtful and insightful with him. He's cool. I like everything, but all my team has. Honestly, I I'll give you an answer soon. What have you changed your mind on in the last 12 months?
1:08:49I used to think that coding harnesses were all like kind of [ __ ] that they don't really like work. Um, and like, oh, you're just going to use like OpenAI. I'm just going to use like Codeex or like Claude. And I'm now like very proh harness like I I see the value in like factory or cognition.
1:09:06And the reason for that mindset change is switching models very quickly and being able to find like the right models for the right task. And so fundamentally it has to be an independent provider.
1:09:15Yeah. Independent provider but also like the tools that they build on top of it just a better experience.
1:09:20So fundamentally the independent harness provider relies on a multimodal world like as a coder and myself and like I use it for like all sorts of other things too. But yeah you can put all of your money into only one company. Which company do you put it in? um today public, private, you name it.
1:09:43Honestly, all it's a this is a wonderful question.
1:09:47Um so I used to tell investors when we started the company and you can ask our current investors this true that I would put I used factory and I would tell our investors put all your money in factory.
1:09:59Um, and I stopped using factory for like 3 months or 4 months because cognition came out and they gave us some tokens and I used it for a while and now I recently switched back to factory better product. I would put all if I had personal cash I'd put it all in factory. You got to get Maton to let me invest.
1:10:15You not invested already?
1:10:17No, I want to invest. Tell Maton I want it.
1:10:22I think the product is so good. I walked around the park with him and I have to admit it was it was one of these or sometimes you meet founders like I did with you where you know so quickly that you're so good but you have to pretend like you're still discussing and you're just like yeah I just so want to invest in your do you know what I mean and you have to play the game you have to play it cool cuz you're like okay great I'm in it's it's too quick no but I want it I I think the product is so good maybe the business is like not like I actually don't know the
1:10:51underlying like the fundamental my business is amazing Okay, the business is doing well and I'm sure they'll sell to like enterprises, but I think the product is so good that I know like a thousand people like me that are probably not using it because they're all like they're using all the other tools that if they use it, they're just going to switch over.
1:11:05Yeah, I'm giving them a free ad, too. They should definitely let me get me in.
1:11:09Um, okay. Uh, final one. No, actually penultimate one. Which of your beliefs about 2030 is most likely to be wrong?
1:11:18Uh, humans will be a biological vessel for AI. I don't think that's going to be true.
1:11:24It's quite bold for 2030 or or that people will live to like 150 years.
1:11:29I I want to believe that it'll be true.
1:11:31I don't think a lot of other people agree that it'll be true, but for 2030 it's a push. I I agree with that. Um final one, what are you most excited for? I like optimism. I think generally the goal of like humanity is to like fundamentally is we want to live as long as possible. That's like our biological goal in life is like live as long as lo long as possible. Live like the best life as possible, like the happiest life. I think AI helps us get there. Um, and so I am like very pro all of like
1:12:00the biotech AI possibilities that exist in like longevity tech especially.
1:12:05Obvious I don't work in that field at all, but I I like to think that the work we do by um help you kind of do you enable people to do amazing work. I I I help I like to think that the work I do by helping people um access compute and hedge their exposure means that there will be more compute in the world and that by having more compute in the world that compute will go to like good causes and will help like humanity flourish.
1:12:30I listen I think you're just an enabler of great things like bio to happen. Yes. Um listen Kush I've so enjoyed doing this. Thank you so much for doing it with me and it's so nice to do it in person.
1:12:39Thank you for having me.