Ellison's Counter Offer, Chinese H200s, Data Centers in Space | Aaron Ginn, Matt Kalish, Emil Michael, Blake Scholl, Naveen Rao, Ofir Ehrlich, Gorkem Yurtseven, Pedro Franceschi

27:25in AI. People familiar with the discussion said, I don't know, it's such an interesting question that I still wrestle with. Like we exported a ton of Teslas to China and BYD and Huawei have now. Arguably completely leapfrogged. Completely leapfrogged. So they did not become addicted to the American electric bar stack. That was the point I was making earlier. Yeah. You can make the art like almost with every single product. They've said, we'll work with you to make this thing that you want to make. We're really good at making things. Yes. and then they ultimately just make a better version of said product and make it for cheaper.

28:08Yes. And in the case of cars, right now we're obviously not allowing these cars to be imported into the U.S., but they can simultaneously say, we're happy to keep making you these things. We're also going to compete with you directly and constantly try to be better than what you do. I just wonder, like, if you play it back and you don't allow Tesla to export the amazing Model S or Model 3, does that slow down BYD's development of their car or Huawei's or Xiaomi's development of their cars? My understanding is they were able to basically get a paid education making Teslas, and they were able to leverage that into making...

28:55But that's about making the car there. If you don't make the car there, which is because the NVIDIA has a Shanghai R &D. But that's research. That's not made there. That's not made there. Like, it's very different when you're saying, okay, we're going to go and produce this product there. And, like, you are going to get educated there. I think it's different. But you remember we've gone through this before. China's, like, five-year plans to create a domestic chip industry. They've been doing five-year plans for 60 years. I know. They're building. That's what I'm saying. and I would argue that it's working.

29:29They're not caught up, but they certainly made a massive amount of progress. They've made more progress than any other country on Earth. Yeah. I don't know. What do you think, Tyler? I think if you want to think about getting the Chinese addicted to our chips, it's like how addictive are the chips? Yeah. Because if you can make a comparison between like you have the NVIDIA chips and then you have the Chinese or like TPU, right? Because the TPU is like in some ways, it's harder to use. The open source is not as good. but if it's just more economical like if the actual hardware is just like a little bit cheaper then it doesn't really matter how worse the software is people will eventually move to it so it's like I don't think there is actually that much soft power in the kind of general open source stuff in regards to Nvidia it doesn't seem very addictive outside of it just being cheaper at the same time we are, America is very much addicted to Chinese solar panels right now.

30:26The Chinese solar panels come here, they're cheap, and so we don't wind up buying or building a domestic solar panel industry because just to get anything off the ground, you have to go in and say, okay, we're going to deal with having no margin forever, and no venture capitalist can underwrite it, and no private equity firm can underwrite it, and so it just doesn't really happen. Yeah, but the Chinese, they do do that. I guess they do bear the cost more, so it doesn't matter. But it's like they wouldn't have to bear the cost, so we're imposing a little bit of a cost on them. It's like they actually have to pay it, whereas they don't.

31:06Whereas they wouldn't if we didn't do that, right? Because if we don't go in and we don't sell them NVIDIA chips, then they're like, okay, well, it makes economic sense to back Huawei and fund them. Whereas now, they're like, it's not economically logical to back Huawei, but we have to because there's so much pressure from better chips flooding the market, like NVIDIA chips, that if we don't fund Huawei uneconomically, if we don't lose money on Huawei, we will never be competitive. Yeah, I think it seems to me that either way they're going to fund it, the Chinese government. I think that's George's point too.

31:48Really, I don't think it actually matters that much. Yeah, I don't know. Well, let's move on. But let's also tell you about Adio, the AI native CRM. Adio builds, scales, and grows your company to the next level. The United States will allow NVIDIA's H200 processors. It's second best. So this is a change from what Ludnick said, right? Ludnick said, we're not selling the best. We're not selling the second best. Turns out we are selling the second best now. Maybe we'll sell the first best. soon. Who knows? Well, we should play this clip from Patrick O'Shaughnessy's fantastic interview with Gavin Baker on Invest Like the Best.

32:28He's talking about space data centers. We had Dalian on the show yesterday talking about space data centers. Dalian still thinks it's pretty far out. Gavin knew we needed a new narrative. He said after the fateful BG2 episode ended the AI trade, we needed a new trade. And now we're getting the space data center trade. Okay, let's hear the space data center. I think the most important thing that's going to happen in the world in the next three to four years is data centers in space. If you think about it from first principles, data centers should be in space. What are the fundamental inputs to running a data center?

33:03They're power and they're cooling. And then there are the chips. The inputs to making the tokens come out of the magic machines. So in space, you can keep a satellite in the sun 24 hours a day. Pretty cool. And the sun is 30 % more intense. And this results in six times more irradiance in outer space than on planet Earth. So you're getting a lot of solar energy. Point number two, because you're in the sun 24 hours a day, you don't need a battery. This is a giant percentage of the cost. So the lowest cost energy available in our solar system is solar energy and space. Second, for cooling. In one of these racks, a majority of the mass and the weight is cooling.

33:47And the cooling in these data centers is incredibly complicated. I mean, the HVAC, the CDUs, the liquid cooling. In space, cooling is free. You just put a radiator on the dark side of the satellite. It's fucking cold. And it's as close to absolute zero as you can get. So all that goes away. And that is a vast amount of cost. And then let's think about how this satellite is going to work. complicated than that what the the dissipating uh heat in space this is what everyone's fighting about i'm not exactly sure uh how it how it plays out like the physics of it but a lot of people are saying like uh yes it's in a vacuum but because it's in a vacuum you can't just you you can't like it's not as passive of cooling it's more of a material science problem like there's more physics uh physics problems there um also like i don't know i feel like you might need some I feel like you might need some batteries because I think there are, there are orbits that are effectively in the sun all the day, but I think that the earth's shadow like hits them, not like every day, but like every once in a while, you'll be, you'll, you'll be caught in a shadow almost in like an, in a, what's it called?

35:05I don't know, some sort of, you know, you just get caught in the shadow not every day but uh like every 12 days or every 30 days or something like that and so in that scenario i mean maybe you maybe it'd be fine if you're doing inference and you don't and you can just turn off the whole system for a little bit and there's some redundancy i just texted our resident spaceman dalian yeah i said answer answer this i just said quick take yeah on this it's urgent it's urgent well well we'll get back from him there there's more to that clip though. We should play this because there's a funny part in here.

35:37How are you going to connect those racks? Well, it's funny. In the data center, the racks are over a certain distance connected with fiber optics. And that just means a laser going through a cable. The only thing faster than a laser going through a fiber optic cable is a laser going through absolute vacuum. So if you can link these satellites in space together using lasers, you actually have a faster and more coherent network than in a data center on earth. Okay. For training, that's going to take a long time. It's just because it's so big, but for inference, so let's think about the user experience.

36:14When, you know, when I asked Grok about you and it gave the nice answer. This is crazy though. They've done a podcast together five times. Why is he asking about Patrick? I was like, what is going on? Went to some sort of metro aggregation facility in New York, probably within, you know, 10 blocks here. There's a small little metro router that's routed those packets to a big XAI data center somewhere. And then the computation was done and it came back over the same path. If the satellites can communicate directly with the phone and Starlink has demonstrated direct to cell capability. This whole thing, I would like to know how much SpaceX exposure Gavin has, because if this does become a dominant, at least narrative over the next few years, a lot of the people that we know that have computers in space are not exactly bullish on data centers in space on a short time horizon.

37:17We have not. We had the founder of StarCloud on. He's obviously bullish, but I don't believe he has any satellites in orbit. We've had Dalian on. He puts things into orbit. He brings it back down. He's talked about having. He technically has computers. I mean, Dalian's a hilarious position because Founders Fund has a huge position in SpaceX, obviously. But also Dalian has had to do the hard work of getting something in space and probably understands how much of a grunt work it is and how much of a grind it is to get even just something as small as the pod into space. And then he's also seen at Crusoe what a real data center looks like.

37:54And the idea of putting something so massive into space. It's bigger than the Hubble. It'd be bigger than the ISS. What are you thinking about the space data center thing? Have you evolved on this? Yeah, I mean, I don't know. To me, it seems just kind of impractical. Like, if you think about if you're doing inference or something, like how much of the cost of the token is actually in the energy? because I feel like one of the main benefits of having the data center in space is that energy is basically free, right? It's right there. But if you think about, I think Casey Hanraar has talked about this.

38:28If you think about an inference, when you're doing inference, the portion of the cost that's actually the energy, it's very tiny because the chip is expensive. It's not that expensive. So it's like, okay, the cost gains overall are not that great, And then it's like, well, what if you want to change the chip to the next generation? Or what if you need to do some kind of like mechanical work? Are we ignoring how a broken cluster would take some sort of intervention to fix as well as the cost per kilogram in fuel costs? Yeah, just for context. So Gavin was at Fidelity and led a$900 million investment in SpaceX at a$12 billion post-money valuation.

39:12That's amazing. Wow. What a goat. that's incredible um okay actually yeah so maybe they didn't do the full 900 million yeah grok isn't sure sure sure sure um but but i mean either way like you can you can be pro or con on space data centers and grok also says spacex has been atreides largest venture capital position as of 2022 comprising an estimated 30 to 40 percent of its vc portfolio yeah feels wild so So anyways, Gavin is set to, on one hand, Gavin could know things about space data centers that the rest of us don't know because of his position and proximity to SpaceX. At the same time, SpaceX, if they go public next year, would benefit a lot from having a space data center kind of like narrative as part of this.

40:03It's crazy because the business is so solid, I feel like you don't need a fourth act or whatever. Okay, wait, so what about this? So if a lot of the gains of a space data center are just because the energy is free, why don't we just, but then there's this hard part about having like different satellites everywhere. Put the data center on the moon. There's no atmosphere, so you don't get the 30 % reduction in energy from the sun. And just put data center on the moon. Well, don't you get a 50 % reduction because there's a dark side of the moon? Like the moon. You wouldn't put it on the dark side of the moon.

40:35No. I know, I know. But what I mean is that our dark side sometimes gets light from when it's not facing us, right? And so if you're on the moon, you only get light 50 % of the time.

40:53Whereas if you're a satellite and you're flying out in space and you're kind of orbiting the earth, but at a weird axis that you're not constantly in shadow all the time, you're getting sun. He says 24 hours a day. I think it's like 99.9 % of the time. But regardless, it's like almost always. Whereas if you're actually like on the moon, you don't get sunlight 100 % of the time. You get it 50 % of the time. I'm pretty sure. I don't think it's because - All right, maybe there's some flaws. Yeah, the sun's not like, or the moon's not sunlocked. It's earthlocked. Pull up this post I put at the bottom.

41:31It's back from Joe Morrison. He's over at Umbra. Well, we pull that up. Let me tell you about Restream. one live stream 30 plus destinations if you want to multi-stream go to restream.com and we will pull up this next post uh joe says you just put a radiator on the dark side of the satellite and he says thermal engineers in absolute shambles right now yeah people are not it's i don't know people are so spicy hammer says to be fair this is a big satellite joe says many big satellites Big satellite makes it easier? I don't know. Yeah, I think you need a lot of mass in order to dissipate enough heat.

42:13From a GB300 or something? GB200? Somehow, AST Space Mobile Inc. is up 27 % in the last five days. So people are quite excited. Yeah. Well, should we head over to Gulfstream news? There's some massive news from Gulfstream Aerospace. The G400 introduces next-gen Gulfstream tech to its class. Sound on to learn more with Gulfstream President Mark Burns. How has Mark Burns not been on the show yet? Here we go. clearly don't have a no render policy over here they're they're render raxing you think they're just not out you think they could just be adding special effects what do you mean this is the most rendered video i've ever seen this looks like it was rendered okay but here is 2010 you think this is real though bring gulf stream performance standards this is not a green screen class fulfilling our customers needs for a product I knew it was too good to be true.

43:19Look at this. You're like, this is real. This is what it looks like. It's far away. It does look like an Ace plane, though. With Gulfstream's signature combination of range, speed, and cabin comfort, the G400 will provide unrivaled efficiency thanks to the combination of the advanced... This is going to be a hot Christmas gift this year. Absolutely. People have been wondering. I think this will be under the tree for a lot of people. it's not ideal to do like a pre-order as a gift you know it's nice to have something under the tree but i think people would make an exception here so funny i was talking i was talking with a mutual friend of ours a guest who's been on the show many times and uh they're planning to purchase a jet at some point in the relatively near future and he was so fixated on like needing to be able to it's his first it will be his first jet he was so fixated on like i need to be able to stand up at, at like, like it can't have, uh, like the cabin needs to be big enough that I can like walk around freely.

44:23I was like, come on for your first jet and you're not just like, you know, uh, tolerate like slightly lower height in the cabin. He was like, no, I gotta be able to, I gotta be able to move freely. Go straight for the 747, the BBJ. Yeah. Get the, the, the VIP, vvip package the one that only the qataris have for some reason uh well uh casey hanmer okay i got it back i got a text back from dallian i won't read all of it okay uh there's a line here this is from someone else dude all this data centers in space stuff also i'm not going to read that you need 1 million 100 kilowatt sats to generate 100 gigawatts in orbit 100 kilowatt sats weighing 1 ,000 kilograms, including payload, would mean a 5x increase in power density, and you can't just not have batteries and slap a radiator on the back, LOL.

45:22And then there was some choice words exchanged that I won't read on the show. But anyways, I think we should host the Data Center in Space debate. That sounds good. We should get a few people on here, somebody to be bullish, somebody to be bearish. I think we want somebody to be bullish that has put things into orbit and taken them down many times. I think it's hard to lean on somebody's kind of take or opinion if they're not actually participating in space. They're just sort of speculating. Yeah. Let's read through a little bit of what Casey Hammer had to say. He said, here's one idea about SpaceX's next big thing, AI computing inference, again, not training, on orbit.

46:09But how the hell can SpaceX do this cheaper than just building more data centers on the ground? From first principles, it's an attractive proposition because the GPUs have extremely high value per kilogram and extreme revenue per kilowatt, both of which are relatively expensive. That is, the value prop somewhat washes out the pain of operating in space. So I took a closer look. If anyone can make this work, it is a Starlink-derived system. So I started with the Starlink V3 satellites with some high-fidelity CEO CAD below. Orbital parameters pick a sun-synchronous orbit. So we're in full 1 ,400 kilowatt per meter squared sunlight at all times.

46:49No need for batteries. Deploy the solar array in sun slicer mode facing full sun. But the edge is pointing in the orbital direction, bottom right in these images, to minimize drag. but the inference Starlink, StarThought satellites don't have to scrape the atmosphere. Being in sun-synchronous orbit, they'll need to use the rest of the Starlink constellation for backhaul via laser links anyway, and higher orbits actually improve worst-case latency very slightly. Too high, though, and SSO is relatively full of debris. Let's pick 560 kilometers. A Starlink satellite in this orbit has full sun, so the back half is always shaded and relatively cool.

47:34The next hottest thing in the sky is the Earth taking up almost half the sky to the bottom left in these images. And so what does he conclude? He says, I've seen a bunch of high inclination Starlink launches from the Vanderburg recently, but I don't think any of them were going to sun-synchronous orbit. In any case, a ring of inference satellites visible at dawn and dusk running north and south will be awesome. If one starship can launch 100 tons to LEO, then that gets close to 30 megawatts of inference per launch. 1 ,000 watts is 300 gigawatts. Now we're talking real scale. We're doing one launch of Falcon 9 per day.

48:12There was Dellians loosely. We're now one a day, so 300 in a year. So you get to 1 ,000 over three years. If you're on a one-a-day cadence or if you can wind up getting more. The math is not that crazy, but it does seem like Delian was pushing more towards like, hey, this is maybe 10 or 20 years, not the next three years. But it certainly will be fun to follow the story. In the meantime, let me tell you about Cognition, the team behind the AI software engineer, Devin. Crush your backlog with your personal AI engineering team. Jordy has survived. I keep getting the... Podcasting a can. Podcasting a can, like down the wrong.

48:53They should really make a drink specifically for podcasters that has no carbonation, that cannot be choked on in any way. Mark Benioff says, LLMs are the new disk drives. Commodity infrastructure you hot swap for whoever's cheapest plus best. The fantasy that the model is the moat just expired. Mark Benioff having fun on the timeline. I love that he's having fun. I love that he's taking shots. Don't they have an AI lab, to be clear? They were working on Einstein for a while, but I think that they are very much happy to be a rapper, happy to be a buyer of LLMs at this point. It certainly has penciled out that way.

49:35We can talk more to Mark about that. He's feeling good. Stock's up 10 % in the last five days. Let's hear it.

49:44Best salesman in history. What do you think about Michael Burry saying the whole industry needs a$500 billion IPO ASAP? So everyone that knows anything knows this. OpenAI is the next Netscape, doomed and hemorrhaging cash. Microsoft is still trying to keep it afloat while keeping it off balance sheet and sucking out the IP. So why do they keep getting funded? The whole industry needs a$500 billion IPO ASAP. The whole industry needs to go away and sit in a corner for a couple of centuries and think about what's done. This is Jeffrey Miller. Wow. I don't know. This is – people want extreme takes.

50:28This post would go super hard if you don't understand the Microsoft OpenAI relationship. Yeah. But, yeah, it's just overly dramatic. What was Netscape's peak revenue? What was Netscape's – I remember correctly. They IPO'd with like$15 million of revenue. That feels a little bit different. Netscape's peak revenue in 1997 was$500 million. I mean, we're up$20 billion this year for OpenAI. Yeah, their total revenue before IPO for the first two operating quarters, Netscape reported total revenue of$16.6 million. Just odd. I mean, did Netscape have like an enterprise business? That's smart. Yeah, I don't know.

51:13I mean, even if there's more commoditization, it's going to take time. And I don't know, the dynamics of the competition here feels like there will still be a lot of value, even if it is somewhat commodity infrastructure. It's like, you know what else is commodity infrastructure? Like AWS, GCP, Azure. Like, you get, you know, like a server with some hard disk on it. Like, that is commodity. And yet, they all have 30 % margins. And they're all massive businesses. And when AWS, when Amazon broke out AWS as its own line, it was like an IPO of its own company because it was so big. It was such a massive business.

51:57And it should be completely commodity because it's just servers in a data center. And yet, it's no margins. Yeah, it's interesting. Yeah, it's interesting, too. I mean, Burry has positioned himself of just hating any company that's overheated, but ChatGPT having close to a billion weekly actives, and ultimately, even if they just compete in search, right, at least it's a multi-trillion dollar opportunity. Whether or not they fully execute against that is another thing. Are you buying the IPO timeline of Cohere, then Anthropic, then OpenAI? Or do you think they will be sorted in a different pattern?

52:47What was the order? Cohere first. That's Aiden Gomez's foundation model lab. He is, of course, on the Transformer paper. Also a dripped-out technology brother. That's right. With a fantastic set of outfits. then Anthropic and then OpenAI OpenAI third that's the rumor I have no idea how the market will react to Cohere yeah Cohere has shifted more into the business fully enterprise not in the timeline at all that could very well be intentional but it would be it'll be interesting how much excitement they can build around the IPO, assuming they're losing a lot of money, assuming the enterprise is really competitive and they're competing with Anthropic and OpenAI and open source models.

53:42If the narrative changes to the point where this idea of every enterprise is going to need basically a custom LLM that's fine-tuned or pre-trained on their data, like what AWS CEO Matt Garman was talking to us about, about these pre-training checkpoints where you can go in and say, okay, I have a business and I absolutely do not want to upload my data to Anthropic. I don't want to update it. I don't want to upload it to OpenAI, but I want a model that understands my business's data at a core level in the pre-training step and Cohere can offer that and show traction there. That feels like that could be the next wave.

54:23We've talked to a number of founders that are building those sort of like custom pipelines and there's not really a public company that is even anywhere in that narrative. And that narrative does feel very nascent. It feels like it has not percolated through the public markets yet. So I don't know. It could be exciting for them. Well, let me tell you about - They've only raised$1.4 billion. I did not realize they raised - Sorry,$1.5. $1.5 billion. $1.54 billion. Very modest amount. Yeah. especially when you have seed stage companies raising half a billion. We got one of those on the show today.

55:02We have Dean Rao joining later. He kicked the hornet's nest on the timeline. What did he do? A little bit. People were pushing back. Because of his racing history? No, no, no. Oh, people show respect. I love that. They show massive respect. No, but I think Jack at Slow was saying that NVIDIA was valued at$500 million when they IPO'd. Of course, that was a very long time ago. But anyways, people were just saying it feels pretty potentially overheated, but we have Naveen coming on the show today, and I'm excited to hear about the opportunity from him directly. I'm ready to say it's overvalued unless he's built a company previously and sold it.

55:47Twice, maybe. Yeah, twice. If he's done that, then he's off the hook. But he's got to prove it to us because if he's just some new grad, some dropout. dropout like a waterloo yeah if he's just if he's just like if he shows up on the show and he's just like yeah i had this idea in my garage i didn't go to school i've never done in business before and now then i'm gonna i'm gonna call him out and i'm gonna say you're overvalued you you raise too much money but if he can prove to us that he's done it once or twice before Then I'll let them off the hook. Yeah, I'll let him off the hook. Anyway, I think that's fair.

56:21Let me tell you about Privy. Privy makes it easy to build on crypto rail securely Spend up white label wallets sign transactions and integrate on-chain infrastructure all through one simple API Let's read this semi analysis post says important a common misconception about open AI's upcoming custom chip is that since it's a custom chip It won't be flexible and will be a data flow machine Open AI has recognized that the 100x efficiency gains for training and inference happen at the algorithm layer and the hardware chip needs to be flexible enough to accommodate these algorithm changes. We went from just pre-training transformers to now doing RL post-training on transformers.

56:57We went from dense transformers to mixture of expert transformers and soon ultra-sparse transformers that will have four active experts per token out of 2 ,048 total experts. We went from casual MHA attention to MQA and GQA to attention sync sliding attention to now even learn sparse attention And despite what AI tourists think, the chip OpenAI is building with Broadcom will be far more flexible than TPUs. Interesting. Despite most of OpenAI's chip team being poached from Google's TPU team. Hmm. The semiconductor horse race continues to fascinate me. And I'm so glad that we have the good folks over at Semi Analysis coming on the show regularly to help us understand it.

57:41I always enjoy talking to them. and we have two folks from Semi Analysis joining the show next week. So we will be having a great time. Let me tell you about Turbo Puffer, serverless vector and full tech search, built from first principles and object storage, fast 10x cheaper and extremely scalable. Of course. That's right. It's scalable. A hedge fund was ordered to pay a bonus to a trader who made 97 % of its revenues. This is hilarious because when I read this at first, I thought it was he had made, he like his target bonus was let's say 10 million and he brought in 9.7 million and so he was 97 percent of the way to his bonus and they were like you didn't you didn't hit your bonus buddy you you don't get the bonus at all so i was like oh that sucks but like that's kind of the deal that kind of makes sense but uh evolution capital management has to pay him because uh yeah so a hedge fund that was sued by a trader for refusing to pay a performance related bonus despite him making 97 % of its revenue, has been ordered to pay him$5.4 million plus interest by the High Court in London.

58:47When I read the headline of the story, I expected it to not be in the mid-seven figures. Would you expect? I mean, I was hoping at least eight. Yeah, he's not getting a G400 off of this. Come on. Not even close. Not even close. Robert Gagliardi sued his former employer, Evolution Capital Management, in London, alleging that it acted in bad faith by denying him a$7.5 million discretionary bonus after he had generated more than$60 million for the firm. Wow. Wait, does this mean that every other trader... Lost money. Something like that. I don't know. Heavily... So he was at the fund, brief stint, he made$12 million.

59:30He was already paid$7 million, including a$625 ,000 signing bonus. Base salary was$425K and a$6 million new issue bonus. So he needs to get out of this. He needs to go into AI research ASAP if he wants to be putting up real numbers. This is rough, but there are some harsh words here, he said. Gagliardi, a block trading specialist, alleged that he was told in early 2021 that a return of$10 million over the rest of the year would be an excellent result. When Gagliardi asked the fund's founder, Michael Lurch, for the payout in 2022, he responded, I'm not going to pay you the bonus if you sue me. So Gagliardi did.

1:00:17And he won. He did, he did. Although Evolution did not dispute his extraordinary returns, the fund argued that the damage done to its reputation as a result of dealing with a U.S. SEC probe into some of Gagliardi's trades that his previous employer outweighed his performance and that the bonus did not need to be paid because it was discretionary. There's no question that Mr. Gagliardi made exceptional profits for the funds, and Mr. Lurch frequently praised his performance in that respect. Had Evolution properly performed the contract, Mr. Gagliardi would have received a discretionary bonus of$5.3 million.

1:00:53The judge also said that Gagliardi was Lurch's prize asset at Evolution. That's a good goal for everyone that is maybe working at a company, become the prize asset. Hopefully that doesn't result in a lawsuit like this. But they're feeling, Gagliardi and his lawyer are feeling very vindicated. Well, we have our next guest here in the Restream waiting room. Let me tell you about Gemini 3 Pro first. Google's most intelligent model yet. State-of-the-art reasoning, next-level vibe coding, and deep multimodal understanding. We have Aaron Ginn, the CEO and co-founder of Hydrohost. Good to see you. It's been too long.

1:01:32How are you doing? Welcome back. I've been all right. Yeah, I just got back from Southeast Asia. Oh, yeah? Yeah, and despite what some of the critics were saying, I wasn't smuggling anything, I should have been smuggling. Accused of smuggling? That is wild. Well, give us the update on your interpretation of the H-200 deal, where things landed, where you think the next debate points will be. I'd love to hear that. Yeah. As you know, I've said a number of times that even though I am a China hawk, that we have to realize that Trump fundamentally is a pragmatist. Sure. And he is not interested in the ideological warfare, and I think that him taking over the Republican Party proves that.

1:02:22And the orientation of reshoring to America, which you all know like I wrote, one of the first arguments for a new Monroe Doctrine. So now you saw that the Department of War has been putting that out. The White House has been very clear that it's trying to reshore its emphasis back to LATAM in South America. and to build up the ring of fire from South Korea down to Singapore, essentially to be more sovereign. Essentially NATO, as we've approached ourselves with having them spend their own money to support their own military, that the expectation is that Asia would do the same thing. So a deal with China on trade was always on the table and was very likely because both parties actually want this to happen.

1:03:03and and i i do believe that with uh like with huawei that that yes it is the national champion and they do want to proliferate it but they are but they're more interested in selling it to other people and promoting themselves as a means of trade like they did with their belt road initiative than they are for internal uses like that that's just kind of my uh estimation talking to people in the region. I mean, it just came back from there. So NVIDIA being in that region and being sold, particularly China, I think is important for American interests. Whether it's an H20 or B30 or H200, I'm kind of less opinionated on.

1:03:46I just think that we need to be there because half of all AI engineers are Chinese. And we can go through any list of anyone scrolling at the bottom of this broadcast. And you can see that probably the majority are Asian descent minimally. And as America, we just have to realize that we're good at certain things and we're not good at certain things. So we're very good at commerce. We're very good at free markets. We're very good at enterprise. We're very good at creating the default framework, the rules of the road for the world to transact. I think when it comes to trying to do these other strategies, I think we're not very good at it.

1:04:22And we just kind of have to realize that also that the president of the United States is interested in making deals. He's not interested in trying to get involved in other people's business. Yeah. Let's I mean, there's like a million different pieces here. But let's start with the fact that there are so many fantastic genius AI researchers in China. Why is it in the why is it in the interest of the United States to help them? Is it that if they start developing on NVIDIA chips, it gets the KUDO ecosystem stronger and we will actually get more open source software basically for free from China that will come back to America and make our AI efforts stronger?

1:05:05Or is it we want them to be familiar with NVIDIA when ultimately they come over here on O1 visas and start working for American companies? I think that's too narrow a focus. It's sort of like Boeing. Do you want Boeing in China? I mean, yes, because I want Boeing to be a thriving American business. I want Boeing to maximize profits. And I don't see the 747 as a piece of – I don't see the 747 as a nuclear bomb. Yeah, I mean, I think that's the key prior on the second one. Do you think that an H-200 is, I don't know, like some form of missile technology? Yes, yes, yes. Increasingly, I have backed off the idea that these chips should be viewed as weapons.

1:05:59Like a defense technology. Yes, yes, yes. I backed off of that. Yeah, because I view it much more like telco technology. I agree. And we did not win at that. We lost that. and that's because we didn't offer the world an alternative. We just were like, hey, don't use ETE, don't use Huawei. And they were like, okay, well, give us something else. And we were like, well, I don't have anything to give you. So what do they do? We're sort of separating things out here because on the one hand, there's give the rest of the world something, and then there's other, which is like give China something, which is like there is a different way, right?

1:06:32Yeah, no, that's true. Because I'm totally down to go belt for belt and road for road all over the globe, right? But the question of like, is there any world where we could get a different cell phone tower deployed in Beijing that's not a Huawei tower? Like, no way, right? This is impossible. But then you go back to the fact of like, you're right, all metaphors fail. That's why they're metaphors. So the other exogenous factors, I think that's different from telcos is the fact of like, the level of AI engineering talent. In fact, it's funding isn't propelling forward American AI. That's just the reality of where we are.

1:07:14The second is the fact of they do not have the founder capacity or the foundry skill set to produce similarly level at scale. Scale absolutely is true. They could be at some point be approaching some form of five nanometer process, and there's rumors around that. So, and that's a little bit of an obsession because it is a barrier of technology that like we, I should say the West has instilled that, not particularly America, Japan and the Dutch. But that's, we don't want to fund that enterprise. And I think that that's where the practice doesn't come into play. It's like if we don't offer them something, they'll just do it themselves.

1:07:51And I'd rather take money from them and build our own sort of reshoring in America. But to be clear, they're still going to do everything themselves, right? No, I don't think so. It's just going to take them. No, I don't think so. You think that China will happily be dependent on foreign powers I mean, last time there were rumors of this sale, they were putting out press releases saying like, hey, if you're going to buy NVIDIA, you've got to prove it. You've got to prove that you need it. You've got to prove that Huawei won't work. It felt like there was a lot of pushback from the CCP. I don't believe at all that they're just going to be like, thank you for the H200s.

1:08:28we're going to stop investment in our domestic chip manufacturing. Yeah. But the problem is in that flow. One, you assume a level of CCP control that's just not realistic. It's a country of like 1.3 billion people with very successful entrepreneurs. Yeah, but they've been ripping five-year plans for over and over and over and over. Yeah, but that stuff doesn't matter. You have to understand that there's lots of propaganda that I think that they say that actually never happens in the real world. Go look at Africa. Go look at all the airports and see actually what actually happened. Go look at the ports.

1:09:05Yes, there are a certain example of ports actually transferred, like the Sri Lanka port. But go look at the rest of the ones. They didn't happen. So they say lots of things. But the question is what matters in terms of the national security orientation of America. And I would much rather take money from their economy and put it back into our economy than trying to give them both the demand side and the supply side argument. You're making a supply-side argument, which I think is very much how the CCP thinks about the world, which is like, oh, we need to build our own chip supply. But in the orientation, I think that, Jordi, you're more biased towards, which is like, well, let's also give them a demand argument.

1:09:38I'm like, no, I don't want to give them a demand argument. I would rather mitigate that as much as possible to say that, yes, they have probably five or six multibillion-dollar operations that want to have these types of chips that are not military-oriented. and by applying these export controls that are overly extreme, you're basically giving them a demand argument as well. And I'm saying, I don't want the demand argument. Like, if they're still going to have a supply argument, that's fine. Like, go buy a BYD car and hopefully it doesn't catch on fire. Like, just because BYD doesn't exist. Are you bearish on BYD?

1:10:12They're catching strays right now. I saw one of them on fire in Hong Kong, no joke. Yeah, I have a video of it. I texted you. Yeah, let's go to the cars and talk about what we were debating earlier. If we're just talking about cars instead of chips, and of course there's going to be a metaphor here, but if we're just talking about cars and we want, and the frame is like we're all China hawks and we're putting on like rah-rah, we want America to win the electric car race. should Tesla have sold Teslas into China because it feels like they did. And then China was able to make the whole bunch of cars that compete directly with Tesla and are, it feels like maybe better, maybe cheaper, maybe better and cheaper on some sort of Pareto frontier.

1:11:04Um, if, if we run that experiment, we try and think through the counterfactual of Tesla had never sold a car in China, where would the Su7 be, the jumping Huawei car, the Xiaomi car, the BYD Han, the one that has the karaoke machine in it? There's some crazy cars. Would those exist? Would they be delayed? Would they happen faster? How do you think about that? So I think this is where there is a reality, which is that they're going to copy. And I don't think that that is something we can change, speaking as a half Chinese American. So it's just the nature of what is the way they think about the world.

1:11:44It's just competition. Yeah, they view competition as poppy. And importantly, there's one world where they need the industrial power of like 1 million H200s. And there's another one where if you give them a single H200, they can tear it down and try and reverse engineer it as much as possible. And then they can also do all the learning curve stuff on their companies, on their national champions on Huawei. I think it is reasonable in terms of an argument against expert controls if you view this as a means of trying to slow down their own model-level progress. I can at least cogently understand what's being presented.

1:12:25I do not agree with this from the perspective of hardware engineering. Because, again, if you don't have the ability to make a car, then what's the risk? Right? Like mitigate their ability to make a car, which is generally been my position. Like I want to mitigate their ability to make chips. And so in a world where there's a demand and supply, in SCCP because they're communists, so they love focusing on supply solutions to everything, which, by the way, doesn't work. I mean we're all free marketers here. We're all a bias towards Austrian. Like we know that their philosophy of the world doesn't work.

1:13:03So they can talk about supply-side engineering all they want, but again, they don't have the ability right now to aggressively address their LeFakery problems. And so in that world, I want to just win because if we remove the demand stuff – because all other companies want to use NVIDIA. If we remove the demand stuff, we're removing one of the biggest parts that would give them justification to accelerate on the SMIC side. But if they're just going to go and play around and be like, hey, we're going to dump money into SMIC because we want to use local, but we're everywhere. It's like the fast food thing of like, sure, they have local fast food, but McDonald's and Starbucks is fresh.

1:13:41And so that's more how I think about it. They're still going to try to copy, but if you don't give them the ability to actually have the toolkit, then it's a much safer transaction. And I think, John, what you said is important is that the debate is highly mixed up between you think an H-200 is an F-35. If you remove that as an A-Prior, then the world becomes significantly simpler to understand. But if you view the H-200, Blackboard, or whatever as a Patriot missile, then there's no real place to have a conversation to be rational. Yeah. Because my view is like I don't have as much of an opinion on GB or B, that kind of like debate.

1:14:27I just – I think American companies need to be in China because we want the ability to not only address the second largest economy in the world and have access to – like that is the Trump position and that's my position. Like we should sell to the world. We should be an export nation. But the other is like China is going to copy, and it's just part of the equation. Flip around for me and give me the, sorry. I wanted to hear shifting to domestic semiconductor supply chains. What is your theory of the current horse race between Tranium, TPU, and NVIDIA? It felt like NVIDIA was dominant last year, the year before.

1:15:13Then we were hearing rumblings of like, hey, maybe AMD is getting it together on the software side. Hey, maybe this TPU thing could be used to train a decent model. The DeepMind team seems to have been able to do that. Are we moving towards more of an oligopoly? What are you thinking about how the American semiconductor AI accelerator horse race is playing out? My view on the TPU side, I don't really know who would buy that as just like a product. I think meta is the closest one that kind of makes sense. But why would any of the cloud companies do it? Because they are in a position right now trying to build their own.

1:15:57And they're trying to deleverage. Why are they building their own? They're trying to deleverage from a video. So why would they want to homingle themselves with an actual direct competitor and sell their stuff? And then there was a... They would be mocked. Yeah. There's also this weird zealotry around TPUs that I made because we had Thanksgiving. thing that's not going to happen. But to say that there's minimal switching costs is absolutely absurd and insane. There's significant switching costs. And so if you're of the mindset of like, hey, well, we have all this capex and it kind of makes sense.

1:16:30And that's justifiable. I can understand that. But that's like the same reasoning that you could apply to the multi-cloud argument. And we didn't really see that happen at all, basically until GPUs. Now you're seeing much more multi-cloud products because GPUs made everyone multi-cloud. So, and if you were multi-cloud before GPUs, you were, as a function of an acquisition, it was not generally intentional. So, I think GPUs make sense for Google just as a, like, vertically integrated strategy because that's what they like to do. And there are probably some other companies that maybe, like, maybe Apple, I could maybe see that because they love low-cost chips just in general.

1:17:10But I don't really see it breaking beyond Google because I just don't know who is really incentivized to buy that. I agree with you on AMD's side, the software's getting better. I think that's just the natural orientation of just LeaseSuit is the most reasonable, because they're competitor to NVIDIA's general purpose computing dominance. I'm pretty bearish on Trinium. TPU is just from a chip perspective I'm bearish on, but as an integrated strategy, it makes sense for them as a company. We're running out of time but I wanted to ask you about one other thing. There seems to be a number of people trying to manufacture a space data center pump.

1:17:53Oh, yeah. As somebody, and even NVIDIA, was posting from their corporate account some renders of a company called StarCloud, some space data centers that they're working on. As somebody that has spent a lot of time in data centers or around data centers all over the world, what are your timelines? Are we talking, Gavin Baker was saying, the most important thing in the next three to four years is data centers in space. I've had plenty of people on the show that believe that it's an exciting possibility or opportunity, but maybe their timelines are more like 10 to 20 years. How do you think about it?

1:18:30Most of the stuff that has been announced will not happen. Ever? No, like, think of it as like a 50%. So if they announce like five gigawatts, probably going to be like two. Right. And, and because this stuff's really building this stuff is actually quite hard and America has not really behaved in a manner to me that believes it's that serious of a, of a thing. Like, this is a very niche issue. And when one reason why the, the going by the H200, the reason why it's making less noise now is because it is, it is a niche issue. It is something that maybe a couple hundred thousand people really care about.

1:19:09and in the data center it affects even though it does bring jobs it does bring uh these types of opportunities to uh cities that previously never had a position to support a data center which is great it is really kind of a narrow set of people that are really benefiting uh and if you go look at uh you can go read the bills that are in consideration across america uh it the america is not acting like it's very serious about building data centers the power profiles the approvals that are happening like it's just not happening versus the other countries they they're generally more serious about it um and many of these announcements that you as as jensen disclosed they're they're they're fundraising announcements they're actually not actual projects yet so so they're trying to let's say they have an equity already lined up but but the equity generally is like 10 to 15 of these projects so you've got to go raise the other portion of it and and the offtake side outside of the hyperscalers, there's not a lot of confidence the credit side has outside of those hyperscalers.

1:20:11So the debt side of the world actually has been quite conservative on these and the private credit. So that's why I never really thought the quote-unquote bubble talk and bursting, we're nowhere really there if you go talk to lenders. We're nowhere even close. The type of capital being deployed has actually been very conservative, has been aligned with contracts has been associated with high paying credit, you know, top tier customers. We're not seeing, you know, the Geordi credit union in the middle of, uh, you know, SoCal doing black, you know, B300 loans. And when that happens, then I'll be like, yeah, you know, yeah.

1:20:48Yeah. We're working on it. Yeah. Like with the, so, so, so, so that that's, that that's where if you look at the legal side, if you look at the power profiles, if you look at what's happening and the political nature of it, I don't think many of these projects actually scale to what they need to. But I do agree with Gavin that we actually do need to get serious about it because it's not like the world will be worse off with more power, more compute capacity. If everything, of all those things get cheaper, our entire life gets better. And just as it's cheaper today to get a banana than it was 100 years ago, and that means there's more bananas.

1:21:25That means people consume more bananas. So the logic will still apply. Let's give it up for bananas. We love bananas. I'll give you a layup for that. I love bananas. Great to see you, Aaron. Software is eating the world. Software is eating bananas. Thank you so much for coming on the show. We shouldn't have booked 15 minutes. We should have booked two hours because it's always a fun chat. Thank you so much for taking the time. We'll talk to you soon. Have a great rest of your day. Merry Christmas. Merry Christmas. Merry Christmas to graphite.dev. Code review for the age of AI. Graphite helps teams on GitHub ship higher quality software faster.

1:22:01What's under the tree this Christmas? It's graphite.dev, baby. It's a great stocking stuffer. Up next, we have Matt Kalish, the co-founder of DraftKings. He's a board member at DraftKings. He's also a board member of Phase Media. He's the founder of Hardscope, which launches today. Let's bring him in to the TBP and Ultradome. How are you doing, Matt? Great to meet you. What's happening? Thanks so much for taking the time. Hey, guys. How's it going today? It's going fantastic. Merry Christmas. Merry Christmas. Please, for those who don't know you already, I'd love to kick off, since this is the first time on your show, with a little bit of your background.

1:22:41You have one of the more interesting backgrounds with DraftKings. How do you tell a story in just a few minutes? well i had a couple friends from some early corporate jobs in my career you know i came up through analytics technology and then over the years moved more into marketing and uh you know i met some of uh you know my two partners jason and paul really at those jobs you know jason at capital one we were early stage business analysts and then uh went to a company called vista print And that's where we learned direct marketing. And, you know, we all had this entrepreneurial bug and we're really looking for the right opportunity to go and spin off and do something kind of in the startup world.

1:23:29Which, you know, we had no experience doing, but, you know, we had that drive. So we looked at a million ideas and then DraftKings kind of stumbled along. We just felt like there was a lot of people that liked predicting things in sports, but didn't really have a great way to do it. So that was back in 2012. We launched the Daily Fantasy Sports Platform. Over the course of a few years, grew that to over 10 million registered users. And it became just kind of a big cultural phenomenon. And then come 2017, that's really when sports betting started opening up in the U.S. So DraftKings was in a really great position to extend our product, launched a sports betting platform first in the U.S.

1:24:16outside of Nevada and became a market leader in that space. And then over the years just kept developing out the product, went public in 2020. And, you know, it's been a really amazing ride. I mean, a growth story that I'm really proud of. Yeah. Take me through. I'm sure there's a ton of questions that we can kick around around prediction markets and sports betting and what's going on there. But first, take us through the launch today. Take us through hard scope. How are you positioning that? Yeah, I had a 14-year run at DraftKings that we've seen it all, the ups and downs, and built a really amazing team over the years.

1:25:00I have the type of team at DraftKings where if I don't come to work for a week, things are amazing and fine. I'm telling you, the caliber, the people, the leadership there, the team has been built out just so well. And, you know, I'm a builder at heart. I like starting things. I like participating in industries that are disruptive, you know, that are in the middle of a lot of displacement change. change. And a couple years back, I worked with one of the original founders of FaZe to buy the company off the public market and bring it back private. And it's kind of a space that I'm just super intrigued by the massive amount of disruption in media moving from traditional forms into creator led to community, this sort of decentralized media through social channels.

1:25:55And And I've also found it really intriguing how content discovery has changed so much, like short form clipping and the TikTok, the shorts. That's how people are really finding content these days. And FaZe is really one of the best in the world at that. So just kind of jumping in, rebuilding the business side of that team and kind of relaunching and bringing FaZe back to number one in creator media. That was, you know, initial focus. And I decided to leave DraftKings and transition out to do this full time to launch Hardscope, which is kind of the platform that FaZe operates on. It's a talent and brand platform to access the creator economy at scale.

1:26:41And it's a space I just think is tremendously underserved and a really massive market opportunity. So, you know, I felt like it was the right time for me to go all in on that. And I'm really excited to be doing it. So walk me through how to actually use the product if you're a brand. Does it feel like just autonomous marketplaces where brands are meeting creators? Yeah, I guess a better question is there's probably been... How much hand-holding is going on? There's been maybe hundreds of attempts at building these sort of brand creator marketplaces. A lot of them haven't. Some have turned into great businesses.

1:27:24There's been little to no venture. Last Friday, Shopmise is sort of in the same space, correct? Yeah, greater commerce tool. But this is more advertising focused. Is that correct?

1:27:42What's special about Hargoop is the impossible part is building the actual valuable content assets that brands want to be a part of and reach the audience. And that's what our team for talent does really well. It's a full stack operating platform. You literally don't need any other resources. We do it all in-house production, socials, clipping, all the distribution side into commercialization like brand partnerships, e-commerce, access to capital and venture for initiatives. and the reputation really of FaZe as the number one creator collective in the world is the proof. I mean, that's a group that operates with this team on this capability set.

1:28:28And so Hardscope is really bringing that to new creator collectives and more verticals, really high potential talent and platforming them to really make the most of, you know, the opportunity that they have, which is really unprecedented. Like the level of power and control over the content, their community and how to commercialize it that talent have today. It's it's a totally new age, really. And some creators have built out great teams, but most really have underrepresented teams in particularly the commercial side. You know, like, how do I take my great content and then maximize the opportunity as basically the CEO of a media business?

1:29:12And so, so, so yeah, let me, I, yeah, I think, I think I was a little bit confused. So you're effectively like a back office for personality, personalities, content creators, whatever you want to call them to help manage everything from ad, ad sales to editing, to clipping content distribution, effectively social media management, like kind of, uh, am I saying that correctly? Yeah, I mean, the landscape is laden with middlemen, let's just say. Like the LA ecosystem is unbelievably brutal. There's middlemen for everything. And we're trying to provide really a full stack solution where you don't need all of this noise to run a really successful business.

1:30:01You need a great business partner, a great team. and we've proven that with the operations of FaZe Clan and how that's come back into prominence as the number one creator collective. I mean, you're talking about a group that had the top seven accounts on Twitch last month that's growing at an unprecedented rate and people thought it was dead two years ago. So like we came in, relaunched the brand and totally turned that around. And that is the proof, that's the proof of concept really that I think is attracting so much interest from other talent and other verticals. And, you know, you had asked about the brand side.

1:30:40That's more my background. You know, I was a large media buyer at DraftKings. We have, you know, as you've probably seen, a lot of advertising out there in the world probably bought every channel you can think of. And the one channel that's hard for everyone is creator direct partnerships. Nobody does this well. And unless you are a pure play specialist where your number one marketing strategy is through creator, you probably are doing this not so well right now. And traditional agencies who try to tack on things like YouTube buying or other creator products tend to not do that very well at all.

1:31:19So we're seeing a lot of opportunity on the brand front to be almost like a new kind of agency of record, right, where we can focus just on creator direct partnership, help kind of bifurcate that channel out from a lot of people just lump it in with audio or other other kind of traditional definitions of channels the opportunity is really breaking out creator direct into its own lane and maybe it's two or three percent of media mix today but the landscape is changing this channel might be 20 30 percent of people's media mix in the future The Unilever said it's over half of their media mix in the next year they're planning.

1:32:00So like a lot of companies are seeing like we are just not reaching Gen Z audiences that are diverse under 30 are just not watching traditional media. And this is the way to reach them. Yeah, it makes makes a lot of sense. How are you? How are you like financing hard scope? Do you have outside investors? Was it, am I hearing it correctly? It was you effectively took phase private and is it the same entity? Or I'm curious how, like how you're planning to scale. Is this something that you want to just operate profitably indefinitely? Or is this, you know, something that you'd one day want to, you know, take public again?

1:32:44Yeah, look, I mean, I've been funding it. I led the acquisition initially, and the biggest reason was it was kind of a mess. I mean, the public company of FaZe was in really rough shape, and it needed a lot of work, and putting in the two years to really rebuild the entire operation. And also to relaunch the brand of FaZe, have that in such a great place today. That was the upfront investment I viewed as getting the ball rolling, kind of building the initial momentum. And I didn't think it made a lot of sense to, you know, involve a ton of people in that. I thought it was something we had to prove as a team here.

1:33:25So there's no noise in the company. There's no outside, you know, investors funding. I do think in this next stage, now that we've launched, as we scale up, there's a lot of interest I have and the team has in certain strategic partners. So I think including people that can be helpful to us. Absolutely. I mean, that's something that we're going to be looking to do in the next year. Makes sense. Little bit of a connection issue, by the way. Sorry. Sorry if there's a lag. We have to ask you about your kind of personal views on the state of both traditional sportsbooks, online sportsbooks, gambling platforms, as well as prediction markets.

1:34:07It's absolutely bloodbath on the timeline right now. Everybody involved seems to hate each other, even the investors on every side of it. And it's an absolutely wild time right now. How do you think this kind of category evolves over the next 12 to 24 months? I mean, I love it. This is one of the most competitive spaces in the world, right? Like you had mentioned with the crater economy, it's very competitive, all these different companies doing this or that. Like, great. I love that. Like, there is nothing better than having a super vibrant competitive marketplace for products, having to compete and win for customer engagement, for loyalty.

1:34:58There's always some curveball in gaming and gambling and sports betting, whatever. something new every year some new competitor some new kind of product variant and uh i think the exchange like how fast that's kind of come in since the election and evolved into a sports product um you know it's it's a fascinating thing and uh i think it'll be very competitive you know draft kings obviously announced they're launching an exchange um pretty much everybody i think in the sports betting space has some plan here. And then there's the Cal sheet poly market, you know, that rivalry. And we just look forward to competing.

1:35:40You know, I feel like DraftKings has proven world-class technology team. We deliver leading products. We have the best customer satisfaction in the market. We pride ourselves on that. So competitions are favorite. it's like something that drives us that makes sense uh what what percentage of uh content targeting men online do you think is entirely dependent on either gaming uh ad revenue it feels like there's huge swaths of the internet that uh honestly like wouldn't really exist or creators who uh i'm curious if you have any kind of internal estimates even like creator media Well, yeah, it just feels like I can think of a lot of...

1:36:27Sucker was on the show talking about, he was saying there's a bubble in sports podcasting, specifically because of all the sponsorships that come from sports books. It's actually made a lot of people in the creator economy be able to go full-time because they've been able to... Yeah, I just think a lot of creators where you look at their business and you're like, okay, 90 % of your revenue is from either some type of online casino or sports books, et cetera. Yeah, it's such a valuable category. Going back to hard scope and expanding beyond just FaZe Clan, which is a sort of mass market, super mainstream media business, going into verticals like sports really aggressively, sports, food, like the content verticals that have a lot of really attractive partnerships, is that's kind of the whole purpose here is to represent the creator economy across all of the verticals at scale for brands.

1:37:29And in sports, you can build out entire content properties and invest in them, knowing that it's going to be supported through some kind of deal, some kind of partnership. It's one of the sure things. So I feel like as a company that builds assets, that makes content from scratch, puts together creator collectives from scratch and scales them. Like we can invent these things knowing that the sports landscape, it's so vibrant that there's going to be deals there. That makes a lot of sense. Like the idea of Benny on food. You said that you're working on food channels. And I think that that's one area of the creator economy that has not been financialized yet.

1:38:11And that feels like fertile ground to me. Obviously business channels. Don't threaten me with a good time. Exactly, exactly. I don't know exactly how it would work. It's like, I'm going to bake this cake. We're going to do a blind taste test. Whoever likes this more, they win. But I'm sure there will be an attempt at financializing baking channels in the future. The Great British Bake Off. I mean, there's probably a prediction market on it already. Will I burn the cookies or not? Will you burn the cookies or not? That's a good question. I don't know. Anyway, thank you so much for coming on the show.

1:38:39This was a lot of fun. Yeah, great to meet you. Congratulations on the launch. And we will talk to you soon. Have a great rest of your day. Merry Christmas. Hey, thanks for having me. Merry Christmas. and, you know, love to come on again in a year and give you some updates on all the great stuff we're up to. That'd be fantastic. That'd be great. Anytime you have hot takes, too, on anything gaming or prediction markets related, I know it's a little tough since you're still on the DraftKings board. You got to keep that in mind. But we appreciate the insight and congrats to the team on the launch.

1:39:07Thanks so much. All right, guys. We'll talk to you soon. Cheers. Goodbye. And let me tell you about public.com, investing for those to take it seriously. They got multi-asset investing. They're trusted by millions. Public.com. In a groundbreaking endeavor, the University of Utah Athletics is entering into what could be a$500 million equity partnership with Outro Capital, featuring the creation and shared ownership of a for-profit entity to operate athletics, sources tell Yahoo Sports. So college football is going to the big leagues. They're going to the private equity leagues. Private equity is now going to own your local college football team.

1:39:48Oh, people absolutely hate this. The problem, I know, I hate it too, because why are they stopping with college football? Please, high school football, middle school, lower school, little league. Preschool. Preschool. Let's financialize that. Let's roll them up. Actually, this is happening. We talked about the hockey rinks that have been sold to private equity rolled up, and then you can't film your own children without paying for the... Somebody out there that really hates private equity just yesterday was saying, you know, private equity has come for everything, but at least you can still go to a college football game.

1:40:26I don't know. And again, so how do we think they're going to, you know, so it sounds like they effectively sold for half a billion dollars. They just sold. You remember it was Chicago, I think, sold all their streets or their parking meters to some sovereign back in the day. It was like for like 75 years. Yeah, so the big thing is that you've got to get the money correct. You've got to get the deal right. Yeah, so they basically buy a college team. You can say that it's some more sophisticated structure than that, but effectively that's what happened. They'll have to immediately put DraftKings or Prediction Market logos all over the helmet and all over the jersey.

1:41:07So that's probably like a$50 million new revenue line that the college could never do. Yeah. What else? So the current mascot for the University of Utah is Swoop, a red-tailed hawk. That's pretty cool. They could change it to draftee. Drafting off the airway. Okay, so another thing, they could increase if you want season tickets to the game when you're a student. Yeah. They could, of course, they're going to jack up the prices. They're probably going to make it like$20 ,000 a year. Let me steal, man. You're being too negative here. $20 ,000 a year, And they're going to offer the private equity firm can bring a private credit partner into the mix and help finance the season tickets.

1:41:49So you don't just graduate with student debt. You actually graduate with game day debt that will stay with you for decades. That's ridiculous. I mean, there are so many reasonable ways to monetize a football team over, you know, the entire life of ownership for this firm. They could sell the windshield. They could sell. Yes, yes. It's very funny to sell everything. But I mean, there is something that actually could be very good for academics at this school, because if you have$500 million in cash, you can go and build a new library, a new science lab. You can go build like actual academic buildings.

1:42:22Yes, you don't really need to. And you don't need to worry about like, okay, we got this donor. Somebody went to the University of Utah. They hit it big. They're a billionaire. They want to donate$500 million. But you know what? They just like football. So they're going to build this new stadium. And you're like, but wait, wasn't the point of this whole thing? unrestricted gifts. Wasn't the point academics? And that might be less of an issue now. So they might be able to fund more great academic activities and endeavors with this money. So I don't know. It's very funny to make fun of. It's very funny to riff on.

1:42:56But there are some potentially really good outcomes that could happen. Anyway, we have our next guest in the Restroom Waiting Room, we have Emil Michael, the Undersecretary of Defense for Research and Engineering of the United States. Emil, good to see you. Merry Christmas. How are you doing? Welcome to the show. I'm doing great. Thanks for having me. Thank you so much for taking the time. You look fantastic. Yeah, I would say thank you for wearing the suit, but unlike a lot of our guests, this is a daily driver, I'm sure. Yes, yes. Well, we were buying a high school football team today, so I had to dress up.

1:43:30Oh, nice. That would be a real treat. Well, you're not. on the ladder. Is it fair to say that you're buying Gemini? Are you buying tokens? Or is this one of those deals where the partnership is structured so that it's not really a major cash infusion for Google or it's not a burden on the taxpayer? It's a little bit more of just two great American institutions working together. How are you framing the deal? it's making to 3 million department of war employees that includes military members and civilians access to Google Gemini on our network which is a private network right so you actually have to port it over and they gave us a great deal for the first year it's like 47 cents so they're very patriotic I'm super excited to have that at the hands of everyone at the Department of War.

1:44:29It's never happened before. Yeah. Were there historical examples that you were pulling from? Because I imagine that service members have been Googling things for a long time. I've heard stories about, I mean, even during the war in the Middle East, the war on terror, they were using Google Maps just to see satellite imagery. And there are a whole host of reasons why a warfighter might want to use a Google product. Is this a new structure? Are you building on the shoulders of giants? Do you feel like when you came in there, the Department of War had a good reference point for how to work with a big tech company like Google?

1:45:07Or was it sort of new territory? How did you think about actually interfacing with Google? I mean, it's entirely new territory, right? Because you could get those apps off the App Store or on a web browser. So to get AI in, there's a lot of sort of fear about AI, what it could do. And then you have to remember when you're doing a Google search, like, yeah, it kind of records where you're searching. But if you're putting stuff in AI, you don't want that to leak back into the model. So you have to architect it in a unique way. So it's kind of new territory. And we did it all in like three months.

1:45:42That's amazing. Yeah. What did the team on the Google side look like? I imagine that they have a large team just for working with the government. You mentioned it was a few weeks of work. But what did they bring to bear to actually deliver this on their side? I mean, they had a ton of engineers because you've got to put it in our networks, which are not simple, right? And you have 3 million people all over the world, all kinds of devices and so on. So they really brought the heat, right? And they had a – for today for launch, they had a room or a war room, if you will, no pun intended, as did we, to make sure that it launched correct.

1:46:29I mean we got a lot of flack on Twitter today. So yeah, I saw these comments, but it was very obvious if you used one brain cell that you needed to be on an actual approved device on a military network if you wanted to access the service. So I didn't get the flack. Yeah. Well, people want to give flack just to give flack, right? Well, I think some people wanted to try it out themselves. And it's like, well, if you want to do that, you can join, become a service member. Yeah, we put a link on the error message a few hours ago for job openings at the Department of War if they want to join so they get access to Gemini from government.

1:47:09We're open for business. Yeah, what else have been the highlights of the year? As you look back on the year, obviously confirmation took time, but then there was this big White House AI action plan. Dan, what are some key moments that you feel like have been maybe under discussed or didn't get enough appreciation at the time? I mean, you know, there have been like three AI executive orders, which is a big deal, right? And Jensen said the other day that no other administration has ever done this much for the AI companies. So we're really championing like these four national champions, right? Google, OpenAI, Anthropic, and XAI, and giving them, like we're talking about data centers on military land.

1:47:59We're talking about how do we get more power to these companies because it's a race. And I think we're the first group in the last six months to really realize that we better get better and stay great and ahead of everyone else on this because it's really the next generation of tech. On the subject of data centers on military bases or land, what kind of comps have you looked at on that side? I know there's like a historically in weapons manufacturing, you have like GoCo's, which seem like, you know, they would provide some framework. But how have you thought about that? We haven't thought about that yet.

1:48:38We haven't broken ground yet or to the business models. We just had the executive order said we're going to do it. and then we got to figure out how do you prioritize who gets what, right? And we may reserve some of that data center capacity just for the government to use for our own purposes. But the point was to send a message that we're all in on AI companies, all in on the data center power chip needs of these companies in a way that the last administration we think was just trying to constrain everything with these orders to really have one company win and that's it so they can control sort of the outputs.

1:49:13we're going the other way. We're going to have all the national champions. We're going to support them all the way. David Sachs is out there making sure the American tech stacks and as many other countries as it can be. And that's a big deal, right? If anyone's going to be using AI, we want them to be on our stack. Yeah. How do you think about the fact that all of the big tech companies seem to be working with the United States government effectively for free, 47 cents, I've heard of deals of a dollar, as high as a dollar. But then I think back to a few decades ago where Microsoft had a deal to develop virtual reality headsets, augmented reality headsets for the US government.

1:49:55And I think it was like a$10 billion deal. And I was really excited when that project went over to Anduril because it's a lot younger, faster moving company. Now, I love Microsoft. I love them as a infrastructure provider. And the fact that they have open AI's IP is amazing. There's a ton of great things that Microsoft does, but developing an AR headset, that always felt like it was going to be a rough go. It was such a big burden on the taxpayer at $10 billion. Much better to see it in a different land, in a different company. How are you thinking about how the government is working with big tech, when to open up the pocketbook and write a billion-dollar deal or when to ask for, you know, 47 cents?

1:50:38Yeah, I think, you know, we want these companies to be successful. So we're going to pay them market rates eventually. It's just how do you get out of the gate? And you get out of the gate by, like, giving 3 million people access and not having to worry about, like, a token count every day, right? And then that's exciting to them because now they get 3 million more customers, right? And they get to learn use cases and so on. And so it kind of works for them, works for us, but we need to pay them fair rates, but they're going to be commercial. Yeah. And I imagine that when a new technology comes up like Gemini, there's just a lot of service members that are just, by no fault of their own, just like accidentally using it a little bit too cavalierly.

1:51:25And this is an initiative to actually refocus on security, refocus on privacy. but can you take me through some of the, you know, I'm not particularly worried about Google having, you know, information from the American warfighter, but I am worried about some of the international chatbots having access to the American warfighters prompts. Can you take me through how the Department of War is thinking about LLMs like DeepSeek, Alibaba, these other AI models? Are those just banned everywhere? How do you think about actually, you know, banning different, you know, vectors of attacks from a security perspective?

1:52:08Yeah. Well, I mean, first of all, we're going to train, you know, everyone is going to train if they want to use it for like, you know, higher risk use cases, right? But this is going to be on their devices, just like the web browser is, and they can still search Google. But when it When it comes to the foreign companies like DeepSeq and Chinese, in the bill that was just released by Congress, I think yesterday, we're going to ban DeepSeq and all these foreign models from use by DOW members and contractors and anyone who touches it because the last thing we want is for those models to get data on how we're using AI, right?

1:52:44That would be a tragic mistake. So I think we're already there, but it's going to be in law pretty soon. Yeah. what's your pitch to talent to come work with the Department of War, specifically in AI? It's like the craziest talent war of our lifetimes, most likely. You're trying to recruit people that can go get, I don't know, anything from a million dollars a year to a hundred million dollars a year to a billion, even up to more. Literally a billion dollars. And so I imagine you have to have a pretty dialed-in pitch to recruit people, and maybe that pitch is you're going to be able to work on.

1:53:27I mean, the H200 news today says that the H200 is not like a F-35. It's not military technology yet, but I'm sure there's a pitch to work with you and your team on actually applied AI in a military context. But how do those conversations go? Yeah, so I call them recruiting Tuesdays, right? I spend all afternoon Tuesdays calling, dialing for dollars, emailing, referencing, interviewing, and my whole team does that. And like the pitches, you know, okay, you have to have some patriotic instinct, obviously. But then maybe you're in between companies or you're just motivated. I mean, Elon has five companies and he still worked through the government for like seven months and brought a ton of those people who were motivated by the mission.

1:54:14So I'm trying to motivate people by, hey, this is the biggest technology deployment in the world. There is no bigger organization than the Department of War. There's no more exquisite, crazy, interesting use cases, whether they're intelligence or fighting, even corporate use cases. And you get to be a part of creating that. It's never been done before. That's a pretty good pitch. And then when you leave the Department of War and you've done that, I think you get more valuable in the private sector, right? All these big AI companies now have big federal businesses. They all have sort of cooperating with the government in different ways.

1:54:49And you have another notch on your belt of good stuff that you've done and been an innovator on, which is pretty rare in government to literally innovate while you're sitting in the job on something that's never been done before that's that big. Yeah. How do you think about the different levels of AI diffusion, AI integration, where AI can actually help the warfighter in the American context? I mean, a lot of people kind of watched the Palantir story evolve over the last few decades, this idea of analytics, just putting dots on a map sometimes, that being enough to identify where an enemy threat might be, for example.

1:55:33How do you think about integrating AI from the most mundane use cases of just speeding up a little bit of paperwork here and there to some of the bigger questions that will ultimately face the warfighters of today? Yeah, I mean imagine like the most simple use cases just like any big organization, an employee in any big organization, right? It's writing PowerPoints for you, writing job descriptions, like making spreadsheets, the basics. Then there's like the cool intelligence use cases. Imagine that we've got decades of satellite imagery, decades of that or sensors that we've had or all kinds of things and now you could, instead of one human analyst having to go So like, I think I see that there.

1:56:16You can go back through 50 years, train a model and say, look for things you've never seen before, right? And then on warfighting, logistics, planning, you know, all kinds of simulations. If you want to simulate a war game in a really incredible way with all the data and all the stuff in there. So it's like pretty compelling, you know, and exquisite to mundane that you could do with this stuff when you deploy it the right way. Yeah. How do you think about this partnership with Google? It feels very much like giving almost the consumer product of Gemini to everyone in the military. How are you thinking about larger projects that might require bringing together custom code, multiple systems, multiple pieces?

1:57:01Because, you know, I mean, just as a consumer, we're running into things where Anthropic might be better for coding and one day and the Nano Banana is better for this and ChatGPT is good for deep research. And if you want to build a business, we see a lot of startups pull three different models off the shelf. Is that something you're starting to look at as like a phase two of AI integration into the military? Yeah, I mean, so ideally we'll have all four models and all four of the newest versions. like all the time at every classification level. Wow. And the reason you want to go up through the classification levels is so you can do more exquisite work, right, more complicated stuff.

1:57:42And then the reason you want all models is just what you said. If you're trying to write code, maybe you want to use Claude if you're trying to do these other things. So as they compete, we want to benefit from it. And we've never had that here before. So it's trying to consumerize it, give choice, but then give more and more capabilities as you go up the chain of, like, of security classification. Yeah. Who owns the data? Does the taxpayer own the data? Does the government own the data? I see a lot of stuff where it's like, oh, this company now owns it. Google's going to have your government data.

1:58:12Google's going to have the data. It seems like they might be storing the data in some points, but they don't have authority over it. How do you frame to like the American who's maybe worried about government overreach or wants to understand where important, critical military data is living? Who has control? Who has final say? Yeah, so taxpayers own the government. Government owns the data on behalf of the taxpayers. Google does not own it. We control it. It doesn't leave our control. And that's why it was complicated to actually launch, right? Because AI models are built so that they learn as you continue to query them, right?

1:58:53And to make it such that the model, we got the latest model, and it took inquiries, but then the learnings from that don't go back into the general model is a form of making sure that the American taxpayer's data doesn't leak anywhere else and just use for the purpose it was intended. How many other organizations globally, allies, things like that, need to take on a project like this? It feels like clearly it's very important to move quickly because if 3 million people can't use the latest AI models in a secure way, it's not super sustainable. They'll eventually go elsewhere to get these products.

1:59:37Do you think a lot of other kind of groups need to pursue a project like this? And are you guys trying to help create kind of a framework so others can benefit? Yeah, I mean, any, well, so we all know, and you guys know, any large organization needs to be using AI just for efficiency purposes, right? And just, you know, dollars and cents. If your competitors, you know, showing more profit in Q1 relative to you because they have too much overhead that it could be solved with AI, you're going to be in trouble. So when you talk about governmental organizations, should every government be using it in some degree?

2:00:10Yeah. I mean, it's sort of like an economic imperative, but it's like a strength imperative too. It's sort of like the way I describe it to people is you're opening the human context window. So So we're not replacing a warfighter, but we're allowing the warfighter to do more with the same amount of time, right? To analyze more intelligence, to do more scenario planning and all that. So I can't imagine any other organization like ours that wouldn't be thinking about this. But hopefully – I think we are setting the pace here. We're going to be the pace setter here for the U.S. government and for governments around the world.

2:00:44I love it. Well, thank you for setting the pace. Thank you also for coming on the show today. This is fantastic. It's fantastic to have you, and thank you for all the work that you're doing. Yes. Good to see you guys. Thanks for having me on. Every citizen. Come back on anytime. Anytime. Always welcome. Thanks a lot, guys. Have a good one. Have a nice meal. Goodbye. Let me tell you about Profound. Get your brand mentioned in ChatGPT. Reach millions of consumers who use AI to discover new products and brands. Patrick Collison had a post that hit the timeline, rocked the timeline with 1.3 million views.

2:01:15He says, two conversations this weekend make me think that there's a vibe shift afoot in Silicon Valley around what one should work on and what is worthwhile. Culturally, it feels like the moment is ripe for new frameworks. One, Davos expert morality is stale and discredited. Let's give it up for the experts. Joe Rogan, Andrew Huberman, Lex Friedman, Tim Ferriss. Those guys don't podcast for Davos. They're not Davos guys. It's very possible the next Dalvos could just be a round table between Joe Rogan, Andrew Huberman. Yes. The real expert. Chris Williamson, Lex Friedman. We love that. That would be fantastic.

2:01:52It's also apparent that the, quote, just be super based counter enlightenment is not really an answer. Yes, woke went too far, but simply inverting it does not work. Good point. Effective altruism is no longer the automatic default for smart people. They pushed it too far. They tried to save too many shrimp from their farming fates, and now they are cooked. Sorry, shrimp. There is increasing skepticism of slot and slop machine dynamics. People don't want to work on a slop machine. They don't want to work on a slop machine. They want to work on something worthy and valuable. The question is, overall, what is worthy and valuable?

2:02:34It feels like this question is becoming more central. What do you think is worthy and valuable? I think it's worthy and valuable that OpenAI hired Slack CEO Denise Dresser as chief revenue officer. It's a scoop. It's happening right now. OpenAI has hired Slack CEO Denise Dresser as chief revenue officer. That really flipped me for a second because they have a CFO. They have a CEO of applications. I don't know where CRO fits in but they got someone and that's very exciting is that gong worthy? I think it's very gong worthy and in other news we got some new SpaceX news from Bloomberg, the chat had it first, thank you guys, we were in the middle of chatting with Emil but SpaceX plans to go public at a one and a half trillion dollar valuation up from the rumored 800 billion just last week.

2:03:34They plan to raise$30 billion, raising, I guess, far above$30 billion. Anyway, so this is reporting SpaceX is moving ahead with plans for an IPO that would seek to raise significantly more than$30 billion in a transaction that would make it the biggest listing of all time. I'll hit this. The Elon-led company is targeting valuation of above$1.5 trillion for the entire company, which would leave SpaceX near the market value that Saudi Aramco established during its record 2019 listing. The oil major raised$29 billion at the time. SpaceX's management and advisors are pursuing a listing as soon as mid to late 2026.

2:04:17The timing of the IPO could change based on market conditions. Let's hope the window stays open, folks. We will do our best to hold it open. I hate when the IPO window is closed. but anyways no comments from the SpaceX team but the junk bond analyst says of course congratulations on saying the biggest number. Always great when you can say the biggest number. Do you think the timing here with Gavin going on about space data centers is just a coincidence or is this part of a allocator conspiracy? I still believe the Primack take that Elon has run the A-B test on public companies. Tesla was miserable.

2:05:09SpaceX was great. And so you want to stay private as long as possible. There's no funding is not drying up for SpaceX by any means. There's no real requirement to go public. There's no reason why he has to take that company public. this i i don't know why stuff is leaking right now it might just be some sort of head fake but it doesn't seem critical uh to me i don't know certainly doesn't seem critical that they go public i would just say you have two major leaks in the span of a week yeah it's just very different than the open ai thing where open ai like in order to win they need to like marshal this capital they need to you know make become gpu rich they need the the gpu prices to fall they They need energy to get cheaper.

2:05:54They need everyone to do everything. They need everyone to get in line. Whereas SpaceX is very much like established. They have their supply chain set up. They've been in this business for 20 years. They're just like cooking and compounding. And I just don't see a need for them to, you know, marshal more capital or do anything really special. It does seem like there's something going on with like, okay, we're talking about a second, third act here. but at the same time it's not like Elon needs to take this company public I don't know do you think he does no a lot of great companies don't need to go public but there's still a lot of reasons to go public yeah yeah yeah it would certainly be exciting and I think it would be a good like it would just be like there's a lot of people that are you know into space and can't get can't get any like allocation into space generally you know they just can't they can't participate in the economic upside.

2:06:49AST Mobile. That's it. That's it. But you can't, you can't, Echo Star is the hold come. I wonder, I wonder, this is, the market's closed now, but I wonder, Echo Star, which owns, I think it's 10, 12 billion of SpaceX stock. Yeah, I wonder what that's doing. Up 6 % today. Well, but it's been fluctuating. So, I mean, back to the Patrick Collison. A lot at the close. Back to the Patrick Collison post, like, what is worthy and what's valuable? Like, I still honestly feel like SpaceX is one of the worthiest and most valuable missions. And, uh, in terms of just, it has no slot machine dynamic. It has, it does not create slop.

2:07:28Uh, it is, I don't know. It's, it seems above enlightenment, counter enlightenment, woke, super base. Like yes, Elon runs it, but just as a company, just the idea of like, let's go to the moon, let's go to space. That should be apolitical. Um, it doesn't, it doesn't even have the Davos expert morality impact on the environment. It's just like one of the, it's one of the purest missions in my opinion. And I feel like it, the, the reason I think there's, there's not, I'd like, there's something interesting about what Patrick is saying here, where there is this highly worthy, highly valuable mission, but it's only, there's only 5 ,000 people that can really play in that game.

2:08:11Maybe like 10 ,000 or something. Like it's just not, it's just not a career path in America. It's not like I'm going into real estate. I'm going into lawyer, legal. I'm going into medicine. I'm going into lawyer. Going into big oil. You can't, you can't just say I'm going into the orbital economy. I'm studying and I'm going to, you know, maybe I'll work at SpaceX, but there's a whole, there's a whole host and like go, you can go into tech. It's very, it's very accessible. You can just like go and get a job in tech. It's very hard to do that in space. and make a career. I don't know. It's nowhere near as easy as it is for tech.

2:08:48Because tech broadly is probably a thousand times bigger. You go into aerospace engineering, and then you have optionality. It's very hard. You try to land a job at a true space company, or you can work the bone. So yes, but what if you're not cut out to actually go do aerospace? What if your actual skill set is marketing? Well, you can go and be a marketer in tech. You can't go and be a marketer in space. It just doesn't happen. They hire like two marketers. Okay. There's no, there's no market. Why is that? So what I'm saying is that like, is that like this question, you're saying we need a federal backstop on marketing jobs in the Oracle economy.

2:09:25No, no, no, no. I just don't see what the problem is. No, no. What I'm saying is that, is that there's a vibe shift because like, like, so Patrick Collison is asking the question, what is worthy? What is valuable? What is a valuable mission. And I'm saying, I know what a valuable mission is. Go into Mars, go into the moon. SpaceX is a valuable mission. But 99 % of people can't participate in that. They can't make that their life's work because if they're not cut out to be aerospace engineers, aerospace designers, you know, if they're not, if they're, if that's not what they're great at, they just can't participate.

2:09:59Whereas when previously in previous areas, like tech was organizing the world's information. Don't be evil. It was just cool. Tech was the mission. Tech was the thing that was worthy and valuable. And so you could show up and say, I'm a finance guy. I got a job in tech. I'm a lawyer. I work in tech. I'm a marketer. I work in tech. No matter what your actual skill set was, you could come to technology, the technology industry, and have a career. And you can't do that in the orbital economy yet. You can't do that in space because there just aren't that many jobs. And so I think there is something that's worthy and valuable.

2:10:34It's just not able to absorb the actual demand for worthy and valuable career paths. And so for a lot of the marketers, for a lot of the lawyers, for a lot of the financiers, they wind up being forced to work in slot machine companies or slop machine companies, because those are the ones that have the jobs. They have 10 ,000 jobs available. I don't know. That's sort of a hot take. Yeah, I think that's fair. I just would, I think you could also make the argument that many people are looking at jobs that say like a Boom Supersonic and they're choosing, they're just making the choice to work at the Infinite Slot Machine company.

2:11:15That's fair. That's fair. Well, hopefully they'll read Patrick Collison's post. They'll say what's worthy and valuable. I got to work for Boom Supersonic. I got to work for Blake Scholl, who's our next guest. so while we bring him in from the restream waiting room let me tell you about get bezel.com shop over 26 000 luxury watches fully authenticated in-house by bezel's team of experts we have blake show look at blake show wow he is in a jet oh yeah he can fly i forgot you can fly right hang on guys i gotta shut the engines down this is crazy what is going on here wow look at It's been a complete rat race to try and put on the most insane performance during a TBPN interview.

2:11:59You did it. I think this is going to take the cake. Blake, how are you doing? Introduce yourself. What's going on today? Hey, guys. It's good to see you. Thank you for having me. It's a big day at Boom. This is a huge day at Boom. I'm so happy. This is amazing. Look at this. Okay, where are we? Take us through this. We are in the Boom Supersonic Factory. This, of course, is the XB1 airplane. This is the airplane that broke the speed of sound in January. Wow. The airplane that resulted in supersonic flight being legal again in the U.S. Amazing. So it's blown wide open. But, you know, a year ago, a year ago, we were joking that it would be way easier to fund this company if we were an AI company.

2:12:41We were laughing. Yes. I didn't think about that on the show. That's right. Yes. It turns out we are. So the engine that we're building, that we've been building for almost four years now to power our Overture supersonic airliner makes the perfect ground power turbine for AI. And so that's today's news. We've got a product called SuperPower, 42 megawatts, natural gas. It's going into data centers. Crusoe is our launch customer. We're gonna be generating tokens and quiet sonic booms. That's amazing. That's amazing. So you guys wanna see the factory? Absolutely. Give us a tour. Give us a tour. So one of the things that anybody building in hard tech learns quickly is that the legacy aerospace supply chain is just really screwed up.

2:13:28And so we are building this factory to go from raw materials in one side of the building and completed jet engines at the other side. So this is some of the raw material that's just come in. This is 17 pH hardened stainless steel. It's heat treated. This is what they call the hot stuff. And, you know, I missed the gym this morning. There you go. Getting it in. Do a little bit here. But this is going to turn into stage five stator vanes that go inside the assembly engine and the superpower gas turbine. So one of the most amazing, surprising things is the technology that is ideal for supersonic passenger flight is actually the same thing you need to power a data center.

2:14:13Yeah. So this is Elon pioneered this at Colossus. Sam's doing the same thing at Stargate. These large arrays of what are called aeroderivative jet engines. And it's like the, they're like the blade servers of the energy world. I put a lot of them in an array. And just the same way blade servers beat mainframes, aeroderivative turbines can beat mainframe gas turbines. So let's talk about how they work. This is a one-third scale model of our engine. It works on a complicated principle. Stuck, squeeze, bang, blow. All right, so let's break that down. Air comes in this way. This thing spins, super high RPM, compresses.

2:14:53You can see the blades get smaller as you go in, getting the air down to about a 20 to 1 compression ratio. Burners in here, this can run on jet fuel. It can also run on natural gas, just a different fuel nozzle. And it goes out the back here. So the model you're looking at here is what powers the airplane. The model that powers AI is very similar. We basically take the fan off the front. We lose these two turbine stages. And then we add what's called a free power turbine on the back. Three stages that extract energy, spin a second shift, second shaft. And that powers a 42 megawatt generator. So like I said, the vision here is we're going from raw materials on one side of the building, completed engines at the other side.

2:15:31The facility we're standing in now is going to do the first 200 megawatts over the next about 18 months. That's pretty significant. And then we're building a much larger factory that's to be able to do two gigawatts a year. We're just going to scale from there. Wow. So let's go inside the shop. Yeah. And this is all on the back of, it was seen as controversial that Boom did not just white label another engine from another company, correct? And now it's sort of come back to benefit you. Is that the correct narrative? I think that's correct. I mean, people called us crazy to not outsource our engine.

2:16:06Frankly, I was a little bit nervous about it. I was just really going to go. It's the best decision we ever made. We're getting a fully custom engine for about a quarter of the development cost. Sure. It enables things like boomless crews. Could not do boomless without our own engine. Sure. It enables a totally new passenger experience. I haven't revealed it yet. Yep. And then I think the most important thing is we can now take that same engine core, put a power turbine on it, and get to profitability years faster than otherwise we could have. This makes, basically, superpower makes not just electricity.

2:16:36It makes capital. and it makes the capital that finances the capital-expensive development of the Overture passenger airliner. So let me walk you around the shop here. So this is basically the first unit of what we will copy-paste into the facility that will do two gigawatts per year. What are Jejuns made of? Well, you've got big round things. And actually, if you look inside this machine here, this is called a turn mill. This machine weighs 65 ,000 pounds, and spinning on it right now is a 4 ,000-pound donut of cast Inconel, and that's a nickel alloy. So this is a relatively hard alloy. This is going to be the turbine center exhaust frame.

2:17:20This is the first very large part that we're machining in-house, so big round things. Then you've also got discs with blades all around them. That's called blisks. These are some of the hardest-to-make parts. And so we have started what we call the Blisk Krieg. And this machine here is actually a more than five-axis mill. So this will hold a disk of forged titanium or forged powder metal super alloy. And then the disk will kind of rotate in and out like this while another cutter head comes down and basically sculpts each individual compressor blade out of metal. So ultimately, we're going to have a whole bunch of these machines cracking out disks.

2:18:05Over here on the other side, this is part of an automated production line for compressor blades. So to give you an example of this, this is what's called the stage one variable guide vanes. This goes in the front part of the compressor section of the engine. These actually move as the engine changes power settings to have the optimal airflow. And this machine here starts with that bar stock, those kind of heavy beams I was showing you outside in the hangar floor. It comes in on this feeder in one side, gets held on the machine. The cutter head comes over, cuts it away. Another gripper comes over, grabs it, machines the backside.

2:18:46A robotic arm comes in, grabs it, puts it on the table, which starts with that story. So this thing is ultimately going to be able to run 24-7 building engine parts. Quick questions about the factory. Where are we? Where is this factory and how big is it overall, square footage roughly? Yeah, so this building is about 70 ,000 square feet. It is five minutes from our engineering HQ in South Denver. And we're about to open early next year another factory that is three times the size of this one. And that's what's going to be able to do two gigawatts a year. Yeah. So what are the key challenges now?

2:19:23Now this feels hard, but maybe more straightforward than a supersonic commercial flight. Like what are you guys doing? Maybe less approvals. Yeah, yeah, exactly. And have like massive, massive demand. Yeah. What are kind of like the key challenges that you and the team are looking out for to be able to deliver on the timeline that you were talking about? Yeah. So after having done a boomless supersonic jet that was safety critical with a pilot on board. This feels like doing it on easy mode. Literally, XP1 had 68 ,000 parts. 68 ,000 parts, and they're all safety critical. The turbine has less than 2 ,000.

2:20:04There's going to be a lot of challenge in getting that to work, getting it to work reliably, getting manufacturing up to scale. After having done XP1, it feels like easy mode. We've got customers that are going to take as much as we can possibly make, as fast as we can make it. Today, we've got the capital to go do that. How much capital? $300 million.

2:20:27Congratulations. Thank you. Thank you. After basically continuously fundraising for about a decade, I can't tell you how good it feels to have raised the round that lets us ship the revenue product that then produces the cash to fund the rest of the stuff. So we're done fundraising. This is the last equity round we ever have to do. Let's go. That's incredible. How was the first night's sleep after you closed the round? Did you? I imagine I imagine that was a pretty, pretty surreal moment, even though the you know, you're still just getting started. I would say people are going to be like a natural concern is like, is this mean we're not getting we're not going to get the supersonic commercial aircraft?

2:21:12You obviously, I haven't lost any faith. This feels like an intelligent move in order to enable that future, basically buy the company time to get us to that point. But what do you have to say to anybody that might be wavering? Yeah, no, this, if you want a super船ic passenger flight to exist, you should be very, very excited about this. The single biggest challenges we had were, how do we prove that we have a reliable engine, and where do we get all the money to do it? and running this thing on the ground proves the engine is reliable and it literally prints the cash that we need to go develop the airplane.

2:21:48So I think this takes us from, you know, less than 50 % chance of success to a far greater chance of success. Like I think overture supersonic flight is basically ineffable at this point. And, you know, some people are saying like, oh, you know, Blake got lost. I'm going to be, you know, I'm just going to become an energy guy. And it's like, guys, I didn't bust my ass for 10 years in order to like only make power turbines. I'm very excited with the power turbine business. America needs it. Like we're losing to China. We really need this. But this is absolutely our bridge to the even bigger opportunity to just reinvent all commercial aviation.

2:22:24It makes a lot of sense. From looking at the existing turbine landscape, a lot of them have supply chain challenges. is part of Boom's Edge that you guys are so used to making all of your own componentry and parts that you basically effectively just need the raw materials and you can make stuff happen. How are you kind of avoiding maybe some of the other delays that legacy manufacturers are experiencing? Yeah, well, I think you named it. Being able to build our own parts is huge. and the room I'm standing in here does not exist at GE, does not exist at Rolls-Royce. In this sort of Jack Welch era of everyone getting focused on return on net assets.

2:23:12By the way, what a random metric. They all sold off their supply chains and they can't make anything anymore. And now we hire like disaffected engineers out of Tratt, out of GE. And they were like, holy Toledo, we can resolve in an hour what used to take us three weeks or three months to do at our last companies because they didn't have access to hardware. And if they ever wanted to change anything, it's really hard to change. That's actually one of the biggest innovations here is what we're doing is we're trying to make the world of atoms more like the world of bits. And you can iterate, you can evolve, you can change.

2:23:46And part of that is about taking software engineers, putting them on a hardware engineering teams and automating design workflows. So that means we can like change digital designs really quickly when we learn. The other piece is how quickly can we turn a part? Like if we take a turbine blade and we send it to a traditional supplier to be made, it's going to take six, maybe nine months for us to get that part back going from digital design to hardware. What that means is if you're the engineer designing it and you get it wrong, like it's really bad. You've set the whole company back six, nine months.

2:24:16So now there's a lot of hand wringing. Now there's analysis paralysis. Now you've got to have layers of managers double checking everything. Now you're really slow. But if you build that part in-house, You can actually turn a turbine blade in 24 hours with a 3D print process, heat, treat, phrase, get it out into the engine. So what we can do is iterate really quickly. So we're liberating engineers to move really fast because if they make a mistake, if they want to do an iteration, we can turn it in days or hours in the same building that they work in. That's super powerful. I see what you did there.

2:24:51It was accidental. I don't know. I love it. I love it. Is there anything that you've been tracking on the regulation side that you think needs to change? We were talking about making boomless crews legal, removing speed limits or high-speed travel. If you were to wave a magic wand, is there anything that you'd change around energy production in America today to accelerate reindustrialization and everything? It's much bigger than energy production. We have a huge problem that I think not enough people are talking about, which was we have a permission-based approach to building, not a freedom-to-innovate approach to building.

2:25:34So if we drove to work, the way we build buildings or permit energy plants, you'd have to go file a plan. You have to list out exactly what turns you're going to take. You're going to promise that you're going to stop at every stop sign, not run through any red lights, always drive the speed limit, and then some bureaucrats are going to sign off on that, and then finally you can drive to work. It's insane. That's the way we build a building. That's the way we build a power plant. I think we need to go from permission-based, which just inserts huge delays and huge costs, to, hey, we're going to have a rule book.

2:26:05We're going to have commonplace rules. And then you can pledge to follow them. And if you break them, then you get fined or you get in trouble or you get your permits taken away. But I think we have to really get rid of the entire concept of market pre-approvals if we want to move fast. It's really holding America up right now. It's the biggest problem with building anything physical. Yeah, yeah. Are you guys hiring at all right now? We're hiring as fast as we can find great people. So in fact... I figured. In fact, I'll plug this if you'll let me. If you go to boomsuperersonic.com slash referral and you send us somebody great.

2:26:44We're looking for great engineers, hardware, software, mechanical, propulsion, everything, technicians, CNC machinists. It's hard to find enough great CNC machinists. Send us somebody. We will send you a free Overture desktop model if we end up hiring that person. I love it. That's very cool. That's very cool. Well, this has been the best hard tech tour we've ever gotten. We've had a number of people attempt what you just did. It always goes poorly. Yeah. There was no lag. It's very high stakes because anything can go wrong with your Wi-Fi or anything. I'm very impressed. This was amazing. And honestly, it was just like chilling, watching, learning.

2:27:23This was really awesome. Normally I have to ask way more questions to get information, but this was really informative. I'm so excited for you and the team put together a great group of investors. What a crazy story. I mean, I'm sure the job's not finished. The book has not been written. But you know I've followed your career for a long time, and you're on an amazing run doing an amazing thing. So we appreciate you taking the time to come talk to us. Appreciate your having me and making this so much fun. Thank you, guys. Have a great rest of your day. Merry Christmas. Cheers to the whole team.

2:27:54Talk to you soon. Merry Christmas. Over and out. Goodbye. Over and out. 8sleep.com. Exceptional sleep without exception. You close the$300 million round. You need to sleep on an 8sleep. Fall asleep faster. Sleep deeper. Wake up energized. If you're trying to build a 42-megawatt natural gas turbine while working on a supersonic jet, But you definitely need a good night's sleep. I want my 8 Sleep to be powered by a natural gas turbine. 100%. I want to – just give me a mini boom turbine for my bedroom, please. Just a small one. It doesn't have to be full scale. Yeah. But just something small that I can really rely on.

2:28:34Yeah. Because – Yeah, it is weird. The 8 Sleep, you plug it into the wall, it's electric. You could get a diesel. A diesel version. or you just fill it up with a little bit of diesel and then you pull start it like a lawnmower. And just leave the window open. Yeah. So there's some air exchange. I mean, the exhaust needs to kind of flow out of the house. Or you could wear a mask and just run a... Well, that could be diesel-powered too. Exactly. You have a diesel-powered sleep apnea mask. That's pulling air out from the outside. That's fresh. Exactly. Exactly. But then you have a diesel motor running in.

2:29:13You can put your whole house on diesel. On your face. Your face is potentially kind of vibrating, but you can see it. Well, people are starting to talk some trash about the old Meta Super Intelligence Lab. Who? 0.005 seconds says, it's now painfully obvious that Meta Super Intelligence Lab went on a massive hiring spree, promised the world, and delivered absolutely nothing. There has been a mass exodus. I feel like Alex Wang was a colossal mishire. When is Zuck going to clean house and admit his mistake? Let him cook. Let him cook. I was the first one to say that Meta Vibes should never have been released or at least advertised publicly.

2:30:01I think it was definitively just not a good product. But let him cook. Let them cook. Yeah, again, the mistake is like people are going to judge the first thing that your new organization releases. It felt rushed. It looked bad in comparison to Sora. But we need to see the next version of Llama. I'm sure they'll call it something else. But give them more than six months. What should they actually do on the product side? Like Google has been able, Gemini I feel like has been able to carve out a really unique position with Nano Banana. and like it went viral and this was Anjane's at Andreessen at the time Anjane Midhaz's take about like the Empire Strikes Back and I have found that like you know having the best video model is a reason for people to go to your app.

2:30:54Having the best you know audio model, the best photo model, the best deep research product. Nano Banana is what the rotisserie chicken is to Costco Nano Banana is a great analogy. great analogy right it's a loss leader it's like come in make a bunch of images we're going to lose a lot of money but we're going to get you hooked on our bread and butter yeah our our language model and so yeah i mean open ai has obviously developed just a great like back and forth chat experience the the voice mode is really dominant there the deep research product it's possible that zuck wanted his true believers to be able to accumulate meta shares at around the one and a half trillion dollar mark because he wants to reward his most loyal potential soldiers potentially uh stammy here says limitless is acquired by meta today end to a lovely journey with rewind and limitless and ryan jones says zuck is going on a generational run of max paranoid missing mobile really really really scared him because he just bought an ai wearable startup he's still spending on the metaverse um i think it's obvious that meta is just going to keep shipping hardware where I think that the limitless acquisition looked like a soft landing for a team that had proven that they can ship products.

2:32:10Again, I know they had customers because some of the customers were mad that Meta acquired them and they were talking about it. So, yeah, it shouldn't be a surprise. I don't think you can – I think it's hard to argue that Meta should just ignore any forward-thinking product lines and just do. Yeah. Well, speaking of the metaverse, we should watch this video of a robot that was clearly being teleoperated absolutely destroying a water bottle. It's one of the craziest videos I've ever seen. This is like so insane. So the operator clearly takes off. It feels like it should be AI or something. Daniel says the hand coming down with enough force.

2:32:56explodes the water bottle, which is not easy to do. If you just have a water bottle and you can't just rip that thing open. But the robot teleoperator takes the VR glasses off and then the robot just falls backwards. Is this real? Imagine having one of these hanging out. It's comedic genius. Imagine having one of these hanging out in your office and it's just running bits all the time. Like its whole goal. I have kind of class clown energy in the office. but if I could outsource that to Optimus and just be able to focus more on my work and Optimus is just going around running bits, I mean that.

2:33:33You know what we have to do? We have to get one of these robots that, you know how so many of the robots are claiming like, we're gonna do your dishes, we're gonna do your dishes. I just, I want to just have, okay, we're gonna all have a nice glass of wine. We're gonna put 50 different glasses, wine glasses of all slightly different shapes on the table and just tell the robot, hey, clean up the wine glasses. Just load them all in the dishwasher. Nothing's, I can barely do that. I can barely do that without breaking. Oh yeah, all the time I'm smashing these things. And you imagine one of these robots just completely decimating the...

2:34:08Sure, I'll unload the dishwasher just in there. Oh no, no, no, please allow me. Allow me to unload your finest crystal stemware. This would be... Allow me to unload your stemware. You discover an edge case where the robot thinks, Well, to unload the dishwasher, I should break every glass and use a vacuum and just vacuum it out. Way easier than just taking them all out at once. Oh, sorry, sir. Would you like me to polish your fine stemware? I'll polish it into dust. Polishes it into dust. Anyway, our next guest is in the Restream Waiting Room. First, let me tell you about adquick.com. Out-of-home advertising made easy and measurable.

2:34:47Plan, buy, and measure out of home with precision. We have Naveen Rao from Unconventional, CEO of Unconventional AI in the Restream Waiting Room. Naveen, how are you doing? Welcome to the show. Thank you. Thanks for having me. Let's jump right into an introduction on yourself, companies you built in the past, and then we'll get into Unconventional. Yeah, so the first company I built in the AI space was called Nirvana. It was actually their first AI chip company. It was back in 2014, before most people knew what machine learning was. Yeah, just a little too early, but you're back. I think it was actually decent timing, to be honest with you.

2:35:27But, you know, I think I sold too early, to be honest. The second company was actually moving more toward the software and algorithmic side called Mosaic ML. Sold that to Databricks back in 2023. And I led AI at Databricks up until just a couple of months ago where I left to start unconventional. Incredible. So, yeah, jump right into it. Talk about, you know, the vision. I feel like you guys, the perfect amount of controversy yesterday because people were latching on to not a seed round at a$500 million valuation, but$500 million of capital invested at the seed round. It was honestly perfect. You want a little bit of spice on the timeline to get attention.

2:36:09And I'm sure there's been a ton of inbound interest, both from the candidate side and future customers. Yeah, I mean, obviously that was somewhat intentional. We wanted to have a big shocking moment. You know, but I think it's actually pretty rational when you think about the opportunity ahead of us. So what we're doing at Unconventional is rethinking how a computer works. You know, the computers that we use today, the fundamental distractions have been around for nearly 80 years. And, you know, it's kind of weird to think about in the tech industry something being around for 80 years, but it really is true.

2:36:43And now we've come to the point where we've pushed that paradigm as far as we can push it. And, you know, fundamental constraints around energy are now hitting us at the global level. Like if we keep scaling AI like we're doing, I mean, I think AI is amazing. I use it all the time. And I think the rest of the world is going to do that, too. We can't get there. We're going to run out of energy to scale these things up because it requires very power hungry chips. So really, we we're saying, can we rethink the paradigm actually from the circuit level up and build something that's vastly more efficient, like a thousand times more efficient than what we've been building?

2:37:17and focus only on AI. We don't need to do accounting software and artillery calculations and all the things that traditional computers do. We'll let that be in the realm of digital machines. But can we build something that's much more efficient for the substrate of AI? And so the opportunity ahead of us, I think, is nearly infinite. And that's why valuation actually can make some rational sense. So what does your supply chain look like or what do you think it will look like in the near future? Yeah, I mean, when we started this project, there were two fundamental constraints I put on it. One is that we have to be able to solve the problem within five years, because this problem is going to hit us in three or four years, and we need to have a solution ready.

2:37:58Then at that point, when we have a solution, you can't have something that's like, I've solved the science problem, now what? We have to be able to manufacture it. So we have to have scalable manufacturing. So we believe we can solve this problem in a big way within five years that actually leverages the silicon ecosystem. So we do want to manufacture it on standard lithography techniques. Oh, interesting. Yeah. ASML could potentially be a partner in the next five years. Like they're not off the table. Yeah, TSMC we were talking with already. You know, I was out in Taiwan just a few weeks ago for this purpose.

2:38:30And, you know, it doesn't mean that's the be-all, end-all. What we're doing is really building a new set of abstractions. A computer is built on a digital abstraction. Can we move away from that and actually move to something that is more amenable to using fundamental dynamics of the substrate? And if we can do that, we can actually open up the world to a whole new set of potential substrates. Silicon is one of them, and maybe there'll be more exotic things. I've talked to folks who are building 3D printed circuits and all kinds of crazy stuff. And so I think the world is going to be very rich in the next 20 years in terms of new capabilities.

2:39:05Right now, we've got to leverage what we have. Yeah. I mean, in terms of leveraging what we have, I feel like there's been movement from, okay, NVIDIA GPUs are bust. Everyone's just doing that. Then we got some serious movement this year from AMD stepping it up. We got TPU. Tranium's looking pretty good. There's like these ASICs are doing well. There's a whole crop of ASIC startups that are saying, hey, we're going to bake the transformer right onto the silicon. They're going with TSMC. Can you help me understand how you were thinking about creating something that's higher performance, but still leaving enough flexibility to be able to actually work with whatever the next algorithmic paradigm looks like?

2:39:55Yeah. I mean, one part of what we're doing is a very deep co-design. So we're not saying here's a transformer and it's immutable. We're not looking at it like that. The way you define the neural network itself is something that we're going to change and actually link to the hardware. And if you look at biology, there is no difference between the definition of the neural network and the physical substrate. They're one and the same. So the dynamics, the physics of those neurons actually gives you the algorithmic richness that you have. So we want to move more in that direction. It's not this abstraction on a digital machine, which is built out of numerics.

2:40:29All the machines that you highlighted just then, which all have a wonderful purpose, by the way, all work on the same fundamental abstraction. They all use bits to represent numbers, and those numbers represent weights that then are manipulated in some way to build a neural network. We're actually talking about building those sort of effective numerics on the physics, on the fundamental properties. Do you have a view on, you know, we were reflecting on this Mark Benioff post earlier in the show. He was kind of making the argument that LLMs are going to commoditize, that, you know, it's like hard disks and you're going to be able to move them around.

2:41:07There's not a lot of value or not a lot of lock in at the, okay, I got this bag of weights layer. And I'm wondering if you have a view on how the, you know, foundation model landscape will evolve over the next five years. because I want to know what your customer mix looks like. But you can kind of tie that together however you like. Yeah, I mean, transformers are something that works really well today. We were able to scale it up. We were able to prove out that we can build synthetic systems that learn. I think that was enormous in the last several years, right? And actually, not just that can learn, but can be useful.

2:41:42So now it's about, okay, well, we can define what useful is. Now we have these quality metrics for what a token is. that actually has opened up a whole new world to me in a sense where I can say, well, as long as I can supply that quality, I don't really care if it's a transformer. I don't really care what it looks like. And if I can do so very cheaply and very fast, that's all I care about. So now we're really getting to this almost pure play supply of intelligence. That's what we want. So now if that's the abstraction, it's not about a transformer. It's about intelligence. And I can define that with some metrics.

2:42:14Okay. Now I can supply that in new ways. And so that's the way we're looking at it. Like, I'm not looking at the world today. I'm looking at the world in four or five years in terms of a product. Yeah. Can you break down more specifically what you expect your first initial customer cohorts to look like? Yeah. I mean, the way we're going to see all of this stuff expressed is really much cheaper per token cost, like 1 ,500th of cost per token or something like that. And I think our cohort is really, at first, we're going to go after data center as the fundamental rollout. But anyone who's using intelligence for an application at that point, they're going to actually have probably strict requirements on what the capabilities of the models are, what the token quality is defined as.

2:43:06So when we have that metric, we can build toward it. We can say, okay, well, GPT-8 or whatever it is at that point gives me this? And is there a way I can recapitulate that in the kind of neural network we're defining? And the answer should be yes. And so we will have kind of strict specs in a way to build toward in terms of what that intelligence token really means. And so it's really our customer cohort will be anyone who's using AI as part of their application. And we want to supply that faster, cheaper, and eventually in a more ubiquitous way. I mean, personally, if I really want to look out there like 10 years, I think robotics are going to be huge.

2:43:44And I think robotics require us to solve this problem of bringing energy down drastically to actually have more processing on the bot itself. That's how we're really going to open up this world of autonomy. And so personally, I think that's where I'm really excited. Maybe that's just the kid in me excited about seeing that future. But I think that's where we're going to go. Yeah, I mean, I know you talked a little bit about the size of the seed round. But can you help me understand a little bit more? $475 million is a ton of money. Is there like a bill that you see yourself paying where you're like, yeah, I'm going to spend$200 million on this?

2:44:20Or is it more like you're just going to be operating at the level of like, okay, we're burning$100 million a year on a lot of top engineers and AI scientists. And like, it's just a big organization early. And you're just kind of jumping to growth stage company scale as fast as possible? Not really that. So I don't anticipate the company becoming huge in terms of people. Call it 80 to 100 people. I've set a state for a while. But much like a Frontier Lab, like a Frontier Lab spends a bunch of money on GPUs to iterate. Like training GPT-5, maybe 30 million bucks, but you got to follow a bunch of dead ends.

2:45:00And you do that with GPU compute. Our version of that is actually building hardware. So we're going to be fabricating chips. We're going to be trying different things. It's very hard to model some of this stuff, actually, numerically. And we're going to do our best to model it, scale up that modeling, actually, on GPUs. But at the end of the day, you've got to build it. You've got to test it. And you've got to see if the whole thing works end to end. And we're going to do that a whole bunch of times. Yeah. How are you planning to actually use existing Gen.AI, kind of the existing Gen.AI stack to accelerate your own development of an alternative to traditional GPUs?

2:45:37Yeah, actually, that's a great question. It's really interesting. As we start clicking into the tools for designing hardware, they're still pretty far behind. Automation of digital logic has been done to some degree as much like coding tools. But on the analog side, this is still relatively new. And we are seeing now companies that are saying, hey, hey, I can explore the design space of different circuit architectures for you using AI. And we said, great, so we want to either work with them or buy them. I don't know, whatever it is, whatever makes sense for us. But yeah, we were trying everything to use the latest techniques to accelerate our exploration of the space.

2:46:15That's the way we're looking at this is, how fast can I iterate? How fast can I find working solutions? That's our main goal right now. It makes sense. What have you learned from racing that you've applied to company building? and what have you learned from company building that you apply to racing oh boy uh i think racing when you start getting into like truly competitive events you start to see that every like tenth of a second matters when you're coming into the pit lane for instance you have a pit lane speed limiter right so i don't think people even know this but when you when you come in you you hit this pit lane speed limiter and when you drop below the pit lane speed limit you hit the throttle, it pegs it at the speed limit.

2:46:57You basically want to optimize that so you don't even lose a tenth when you come into the pit lane. You basically hit the brakes at a certain point, slow down, and you're at this pit lane speed. You don't want to slow down ahead of time and ease your way in. Every tenth of a second matters. That may be in even a 24-hour race. I think that's true with a company. Don't waste time. Every moment matters. You may make bad decisions. That's okay. Figure out how to back them out. Move as fast as you can. If you hired the wrong person, fix it. And so I think the main thing I learned is that time is your biggest enemy always.

2:47:29I love it. Well, we normally ring the gong for$475 million, but we should also ring the gong for 265 points in the Ferrari challenge. There you go. Yes. I'll take it. They're both massive accomplishments, but we're big fans of the Ferrari challenge around here. So it's a big deal for us. Thank you so much. Imsa is where it's at really. That's the big stuff. Oh, cool, cool. Yeah. Yeah, we... We were out in thermal a couple weeks ago, and it was the first time for both of us on the track, and it was life-changing. It was like a new... And I'm sure it was somebody that you probably shared a track with This is the hobby.

2:48:04Here and there. Awesome. Anyways, congrats on the announcement and everything, and the whole team. Sounds like an incredible place to go work right now. Yeah. I'm sure... I can't wait to have you back on. And what a murderer's row. You got Andreessen, Lightspeed, Sequoia, Lux, DC, VC, Future Ventures, Jeff Bezos, Databricks, and many others. What a fantastic way to kick off a new business. Very cool. Congratulations. Awesome. Thanks so much. Appreciate it. Have a great rest of your day. We'll talk to you soon. Cheers. You too. Goodbye. Let me tell you about Wander.com. Book a Wander with inspiring views, hotel-grade amenities, dreamy beds, top-tier cleaning, and 24-7 concierge service.

2:48:44It's a vacation home, but better, folks. That's right. Our next guest is already in the Restream waiting room. We have the founder of Eon. Welcome to the stream. How are you doing? I'm doing great. Thank you very much for having me. I'm very excited to be here. Thanks for being here. We're very excited as well. We wanted to get you on last week, but we were on the road. And yeah, very excited to chat. Thank you very much. Introduce yourself. Introduce the company. We'd love to get how you're positioning it. And then I want to go into the funding round, some of the implications of this. But kick it off with a little bit of an introduction for us.

2:49:22So sounds great. As always, I always start by saying my name is Ophir. I'm married to wonderful Julia and father to Rome. Objectively, cutest baby ever born. Wow, that's a great intro. I love it. These are the important stuff. Less important stuff. I started a few companies before. Actually, four companies to be exact. Wow. Most recent one was a company called Cloud and Do, which we sold to Amazon on January of 2019. grew the business to a billion dollar business doing migration and disaster recovery to the cloud. Let's give it up for migration and disaster recovery in the cloud. Doesn't get enough love.

2:49:57Doesn't get enough credit. It does not get enough credit. That's right. I agree. I'm actually this crazy guy who really loves data, data regression, data protection. We lived this all so long. And both me and my co-founders, and while we were there, we saw an opportunity that we just couldn't resist all those companies really moving to the cloud, especially around COVID and later on. And all the incumbents in our space were born on-premises. So we decided to start a company in cloud backup. You could say, cloud backup, isn't it solverly like in 1981? Yeah. Well, we figured out that as companies move to the cloud and become very significant, they have a very big problem actually managing the data, Imagine the secondary storage, backups like tapes.

2:50:45These are companies storing their most precious data for a very long time, and they can't access that. So we started as a backup company. I said, I'm this crazy person starting a non-AI company in an AI world. But then you know what? It hit me. AI is such a strong, compelling event, which has strong tailwinds that we just had. I mean, look at the Seagate and Western Digital stocks. They're through the roof this year. It's crazy. And it makes sense because everyone's going to need to store a lot more data. Everyone's creating so much more data. So much data is created all the time. And you always need data.

2:51:21And I don't know who's going to win the compute wars. Yeah, yeah. OpenAI, Anthropic, Databricks, Snowflake, any area. What I do know, no matter what's going to happen, data is going to be the enterprise mode. And we are here to help them unlock this data. We're automating cloud backups. Sounds boring, right? making it into useful, accessible assets, data leaks. So you can backup all your data in a very seamless, easy way, but then use it, for example, to train models. That's all the hype, right? And analytics, NBI. So no tapes? It's on drives? No tapes. As it seems, we're a software-only solution, only in the cloud, cloud-born.

2:52:06This is what we've been doing. Yeah. On a public cloud or on your own cloud? Because I imagine that one of the benefits of backups and one of the key things that people want to buy when they're backing up data is, hey, I'm already tied to AWS. I don't want everything on AWS. I want security. In case AWS goes down, I want to back up somewhere else. How are you thinking about that? Perfect. And we've seen quite a few outages in the last few weeks, right? And what we're saying with democratizing your data, it's your data. Whether it's AWS, Azure, and GCP, great clouds or other clouds, we're making sure you can extract the data, make it portable between clouds, make it accessible, performant, browsable, and searchable.

2:52:48You can search and browse all of your data and all of your history immediately. And because we're building our own software stack, software only, it's not going to cost you a dime. It's actually cheaper than building it on your own or using existing cloud technologies to do that. So we're trying to make sure that we make it, we give you an easy button as a customer to just enable the use of all of your data, making it accessible to if you want to run Snowflake or Databricks or Redshift or BigQuery or anything that you want on your data, that's great. We are here to make sure that you're enabled to do it on all of the data that you have and all of the data that you had in the past.

2:53:28But how can you possibly be cheaper than the big hyperscalers, the big public clouds? I imagine that Microsoft, GCP, Azure, AWS, they're buying so many C-Gate drives, so many Western digital drives. How can you possibly compete on price? That's an unbelievable question. Actually, I don't want to just win on price. I just want to make it easy for you. So what we're doing, even though we're still in the cloud, using cloud building blocks, object storage, GCS, S3, Azure Blobs, etc. Got it. We're building in a very different way and actually building a more efficient version of cloud snapshots that purposely built for backup and data retrieval.

2:54:11Okay. Yeah, that makes a lot of sense. Take us through the news. What's the funding news? I want to ring the gong for a data storage company. Thank you very much. I'm also very excited. So we raised$500 million in total,$300 million just in this round. Really excited. Led by Elad Gil, who is not only an amazing investor, he's also an amazing human being. What's it like being a fly on the wall when you and Elad get together and talk about data? Well, Elad is simply unbelievable. You know, he's one of the most humble people I've met. and at the same time, brilliant, brilliant, and so well-connected.

2:54:55I don't think I've ever seen an investor that everyone likes so much. And I have the privilege to choose amazing investors. I have Sequoia and Lightspeed and Greenox and Bond Capital and the likes of Omri Caspi who are just amazing, amazing investors. And Elad Gil is very unique. Actually, we got preempted by 16 VCs and I so wanted it to be a lad, really. And when a lad wanted to, I'm going to lead my round, I was in New York. He flew from San Francisco. He came to my office. He basically wouldn't leave until I signed. And to be honest, I really, really want to do that because I got to know the person, not just the investor, in the last few months and a year, he put a small chick in a previous round.

2:55:44And honestly, we just fell in love, not just myself, my other co-founders as well so that's incredible when we had the chance we had to do it and and really really happy excited about it you wrote the book on high growth I know I know and definitely and Absolutely, absolutely. And I'm really thrilled. We have now a lot of capital in a time when we can now plan ahead. We don't need to think about the macro. We don't need to think about the macro. time. This is hilarious, but it's a great take. It's a great way to build your company and your business so that you can take advantage of these tailwinds, but also survive the headwinds.

2:56:39And congratulations to you. Absolutely. This is what we need. Exactly. We're here for a long time. Yeah. Forever. Forever. You've done four companies prior to this. Do you want this to be the last one? You just want to just... Last one. Last one. All the way to the IPO and beyond. Yes. that's my goal fantastic incredible you have incredible energy I'm excited to follow follow your journey and you're welcome on the show anytime thank you so much thank you very much I'm really excited I love your show amazing thank you very much I'm going to get back and tune to that thank you thank you we'll talk to you soon have a great rest of your day cheers goodbye that is worthy and valuable we were debating what is worthy and valuable that feels like a company that's the worthy and valuable company of San Francisco.

2:57:27He's keeping... That's what you need to do. I'm glad he's keeping all of our data backed up. Yeah, yeah. It seems extremely valuable. He's that guy. Well, we have Gorkum from fall in the rest of your waiting room with some massive news. He's back. He's back on the show. You know him. You love him. Good to be back. He makes all of this possible. Every time you're on, I'm like, he's going to be back on. He's going to be back on soon. How you doing? On a tear. Merry Christmas. Merry Christmas. Thank you so much. Good to be back in the temple of technology, capital of capital. There was one more. The fortress.

2:58:05The fortress of finance. The fortress of finance. There you go. Thank you so much for coming on the show. Give us the news. What happened? Yeah, today we announced our$140 million Series D. It's led by Sequoia and meaningful participation from Kleiner Perkins and NVIDIA. It's the Sequoia venture capital firms. We had an incredible year this year. We have some of the biggest advertisers, retail platforms, design and productivity apps, and movie studios generating images and videos on the platform. And we 8x star revenue in the whole year. So this is our third fundraiser of the year. We announced our Series B earlier this year in the beginning.

2:58:55You got one more? You going for the four, Pete? One a quarter? Come on. Come on. We got two weeks. We got two weeks. A couple weeks. What's been the biggest driver of revenue growth? Has it been uptime, being a one-stop shop? Like what messaging has really resonated the most and driven the most growth for you this year? Yeah. So beginning of this year, we thought we were going to ride the AI video wave for the whole year. That was accurate. It was an incredible year for AI video. But what was surprising to us is actually rise of image editing. Like image models existed a couple years ago. like that's how we started our business.

2:59:43But image editing, one of the first good models came out around May. And now it's a bigger part of our business than AI video. That was surprising to us. So we thought we were going to ride a big wave, but another even bigger wave, you know, collided with it. And our business grew with it. Well, is that surprising? Because, I mean, when you think, if it's very commonplace now to like one shot a great image output. And it's so much harder to one shot or even get a great video output. You can get a decent video output. So I feel like it makes sense. Everybody's making images all the time. It's like, it has, it's like, when you think about in the workplace, there's just so many different use cases, even, you know, for consumers.

3:00:31So I think it makes sense. And it really is just exciting because the image models feel like they're so good now. everybody's so used to seeing it and getting kind of like faked out by an image now because you can't tell the difference video you can still usually tell but it feels like maybe in a year from now it'll be the same situation where it's like I just don't know what's real I don't know what's what's fake it's it's all confusing 100 % video video models are still very hard to work with still some really talented I could creative people create great content with it, but it hasn't really reached mainstream.

3:01:07I would say it takes even more effort to create an ad with a video if it's like high quality than actually shooting it yourself. But what happened with image models is going to happen with video and the user experience of video models are going to get much, much better and you'll have even more mainstream adoption of these models. I'm pretty sure of this. And that's what makes me really excited because we went so far. it's a lot of image models, a lot of image editing, just a glimpse of video models. And this is all gonna get better and better. And it's gonna reach complete mainstream. All the studios, all the retailers are gonna make use of this technology.

3:01:49How, oh, sorry. Yeah, just what are you seeing? Like, how are you thinking about 2026 from a model progress standpoint? you know, Nano Banana was obviously, you know, massive leap. I'm expecting we'll see even more activity in 2026. But I'm sure you have somewhat of a preview just given your guys position in the market. Yeah, exactly. So usually, one of the frontier labs pushes the boundaries. In this case, it was Nano Banana twice this year, it was like their first release. and now this very recent Nano Banana Pro release. But very quickly, either the open source community or labs from China, people catch up.

3:02:38There are a couple of reasons for it, to be honest. Once they see there is demand for it, once people know exactly what to build for, it's usually easier for them to get motivated and catch up. But also some of the tricks, some of the research tricks leak and people use them and train these models. The model market for generative media is a lot more fragmented than LLMs. There's so much choice. There's so many models that are different and that are better at different things that the best model keeps changing. Even Nano Banana was dethroned a couple of times throughout the year. Right now, it's considered the best image editing model, but I don't know how long that's going to last.

3:03:22What are you excited for in terms of capabilities for next year? Do you have specific benchmarks that you're tracking or even functionality? I've noticed I have one that's the Where's Waldo test where I will ask you to generate a full Where's Waldo. And even Nano Banana Pro still can't quite do it. It'll either put Waldo right at the center or it'll make two Waldos. And those images in the children's book are really complicated. It's not just a portrait. And there's maybe not enough training data on the internet. But that's one test that I've been tracking. Do you have your own internal benchmark or whatever you think is next, the thing it can't do right now?

3:04:09Yeah, I have some of my favorite prompts that I try with every single model. But I rely on our team. We have great creative people in the team. They spend all their living hours working on these models. They are usually ones coming up with very creative ways to utilize. One thing that changed with Nona Banana is you can actually feel the model has more world knowledge than other models. And that opens up other doors. I don't know if you've seen people are now adding some web search and creating newspaper articles on the fly, things like that, that are very creative uses of these models. And on the video side, people really want more controllability.

3:04:55They want character consistency. They want scene-to-scene consistency. And some of the releases that happened recently, Kling, for example, is one of the better video models out there. they've announced some editing capabilities which are pretty remarkable and all of that is going to get easier to use and once people have more consistency with the scenes they're creating with video models i think that's going to make a big difference yeah i always see those videos of it's a little yachty walkout uh that is like character swapping and it's always really obvious what's going on it's rough around the edges but it's still hilarious because it's a great meme and I definitely would predict 2026 is the year that that little Yachty walkout just looks 100 % real.

3:05:46Next fundraise announcement. I can't wait. Yeah, yeah, yeah, yeah. We need to do that for you. We need to get your body scanned a little bit. You know, you've never asked this question but you might know the answer. You know why it's called Nano Banana Pro? Where, do you have any idea what the origin is of the names? So they had a code name, Nana Banana. Usually when they put these models into the benchmarking websites, they usually had a code name. Like they were Blueberry was one code name, Red Panda was another, and this was Nana Banana and people recognized the model like that and they kept it.

3:06:25That's the story I know. I don't know if there's another version of the story. But I guess the other question is, obviously it is a codename, but where did the codename come from? Google has a long history of using fruit, although so does OpenAI with strawberry. But all the Android iterations were like ice creams and desserts for a long time. But the question is, what's the nano doing there? Because I want the full banana. Give me the nano banana. Give me the giga banana. I want the giga banana, the biggest exa banana, the mega banana. I want the biggest banana model. Don't give me the shrunken down nano one.

3:07:08I want the giga banana. Yeah, I think it's using the smaller Gemini model as the LLM part. As the reasoning. That's why it's nano. But it's funny to have fun with. Anyway, thank you so much for coming on the show. Congratulations on the massive news. And thank you for supporting us all year long. And you got to, if you 8X again next year, you got to 8X the number of fundraisers too. That put you at like two fundraisers a month, two fundraisers a month, 24 for 2026. We'll see. That'd be great. But congratulations to the whole team. And we're very excited for you guys. Have a good one. Great to see you.

3:07:46Cheers. Goodbye. Well, we have our next guest already in the Restream waiting room. We have Pedro from BRAX, the founder and CEO of BRAX. announcing a massive partnership. We're very excited to welcome him to the TBPN Ultra Dome. Welcome to the show. Great to see you, Pedro. How are you doing? How's your day? Merry Christmas. I'm great. How are you guys? Thanks for having us. We're great. We're in the Christmas spirit. We are in the Christmas spirit. We got bells. Very holiday-themed. Holiday-themed shows for the rest of the year. Yes. But it's great to meet. It's great to have you on the show.

3:08:22Yeah, please take us through the news today. oddly, I'm actually a fifth, third customer. So I want to know how this applies to me. Yeah, you get a Brax cartoon. No, so in all seriousness, thanks for having me. Of course. So today we're announcing a pretty big partnership between Brax and Fifth Third Bank. It's a$5.6 billion commercial card partnerships where effectively, wow. Massive. So good. I heard a big number. I had to say. We love big numbers. I want one of that in my office. Fantastic. And we're really excited just because when you look across all the U.S. and we just see tens of thousands of businesses out there, 98 % of businesses are still using legacy corporate cards.

3:09:05And there's so much opportunity in just automating how these teams run their finances. And Fifth Door just has this massive distribution and trust. And, of course, Brex has the financial services and the software and the AI that just helps these businesses automate so much of their manual work and better allocate capital. And they're really excited for it. Yes, sorry to interrupt. I'm curious when you guys started working on embedded products, because if you go back to 2021 and 2022, I'm sure there was a bunch of companies that, or at least a few that got created to go and pitch some of these same banks and say, hey, your corporate, I don't know if people remember, anybody that remembers SVB, like corporate cards back in the day.

3:09:45I had a few of those. Their UI was truly insane. It was like, how is this Silicon Valley Bank? And you had like a different login too. It was crazy. It was truly the most heinous product experience of all time. And it was like, this is Silicon Valley Bank. We're supposed to be the tech. But yeah. Beat a dead horse there. Yeah. Kind of a dark moment. At some point, you said like, hey, we'll go compete here as well. Yeah. Yeah. So for us, it's really been driven by a lot of just customer demand. So we started this with Navon and Zip, where they really saw a lot of the value in bringing financial services and card workflows into their software.

3:10:22Sure. And we're really excited about it because it's just a one plus one equals five situation. Both products become better together. And then as we thought about the rest of the U.S., and 8 % of businesses are now accessible through this partnership, for example, in the U.S., which is a really exciting thing in the commercial space. And the really big thing is how do we make the bank's product mysteriously better? And when we bring the technology and what we've done on the card side, and especially the financial infrastructure that we've built from scratch, like we don't run on Stripe issuing or Marketa.

3:10:50We just build our own rails from the ground up that enables a partnership to actually take place. So it's how does it how does it actually work? Is it is it fully white labeled or does somebody in First Third get kicked out? No, no, no. It's actually it's actually a fully a fully Brex branded card. And the reason is because Fifth Third wants to have someone that's actually responsible for delivering this great customer experience on the card and implementing the software and rolling that out to their customers. and what they do really well is the banking side, right? It's the lending side, it's the banking, it's the relationships and the thousands of branches and just the thousands of distribution partners that they have on their side.

3:11:31And the tens of thousands of customers they serve are all effectively reliant on that and already have that relationship. And now they get a BrexCard, which we're really excited. Yeah. Sorry, you mentioned you're not using Stripe for, what was it? Issuing. Issuing. doing um does does stripe or do are there other fintechs that do play a role in like your supply chain or are you like down at like an ach layer like just like what what's kind of going on under the hood yeah so what we've done since uh 2018 umbrex is we actually build our own financial infrastructure from scratch okay so we actually went straight to the metal mastercard and built everything from the ground up and because we vertically integrated the whole thing sure this created this pretty big advantage on everything that requires their metal capabilities on financial services.

3:12:21For example, global. We operate in over 200 countries today with local cards and local issuance because for us it's just a matter of saying, hey MasterCard, we're going to issue in Europe or Brazil or India or Mexico and just flipping a switch for them because we are our own issuer. We run our own stack. And then Fifth Third was another example where we said Fifth Third wanted to issue a Fifth Third issued card. They want it to be their name on the card. It has to be under their license on MasterCard and all of that. And for us, it's just effectively bringing in a new bank live, which we have a few banks on the back end, including Fifth Third, of course.

3:12:55So for us, it was sort of piggybacking back on this advantage that we've had over many years by building our own rails from the ground up, which is really... How are you thinking about... We've had a bunch of different teams and companies on the show building new layer one blockchains for payments. They're touting like speed, default global, all these things. We've had a bunch of stable coin issuers and business. I'm sure in conversations that you have, you know, you're being pitched these companies to be an infrastructure provider. I'm sure your investors are like maybe asking you sometimes, should we, should there be a Brex stable coin?

3:13:33You know, who knows, right? I'm sure these questions are coming up. How do you think about kind of the layer one landscape and you're operating a global business today. And so I would, you know, your opinion on some of this stuff in terms of the real value of it, I feel like would carry a lot of weight. Yeah. So the way we think about it is maybe a little bit different, which is we think that first like stable coins and blockchain in general have a tremendous application and we're really excited about a lot of the work there. We announced support for stable coins this year when it comes to our banking product and you can pay the car to stablecoins.

3:14:07We've done a lot of that. But when it comes to global, right, and just supporting customers in a wide variety of countries, really the challenge there is the on-ramp and off-ramp, right? And specifically, you know, the fiat last layer for, you know, an employee to reimburse you in, you know, Indian rupees or Brazilian reais and then for you to settle the card in that currency and do all that locally. And unfortunately, stables don't help you that much on that last layer. And the whole, actually, a misconception about the way a global product works is a lot of the value comes from remaining in local currency because what they're trying to do is minimize FX.

3:14:45Even on stables, FX has a cost just intrinsic to it. And a lot of what we're trying to do is keep you entirely in reais. So your revenue in reais in Brazil are settling, your expenses in reais or in euros or in GBP or whatever currency you're in. So a lot of the infrastructure we build to enable this multi-currency and multi-entity operations versus having to do the effects and a lot of the benefits that stables typically bring to the table. So our use case is a little bit different. I'm sure that's disappointing to people building global payments all once. They're like, wait, you're supposed to need this.

3:15:19I mean, we do use it internally for some stuff, but I would say the value prop for our customers is we stay in local currency as much as possible because it just minimizes so much savings and effects. What are the other keys to success? Obviously, you have experience internationally, but you're like a YC company. I think of Brex very much as like an American company. But then in terms of winning internationally, what does it take? What have the lessons been? What are the interesting insights or what were some surprises that cropped up on the journey? Yeah, so I would say the biggest one is that when you go into like true global, like true enterprise, the amount of requirements and specifically on both the card and expense management side is unbelievable.

3:16:10So you have these rules, for example, local per diems. So when you're in Europe, there's these rules that if you have a German employee traveling to London, there's one set of rules. If you have a London employee traveling to Germany, completely different set of rules. And then if they rent a car, there's different mileage requirements. And then all this information is something you have to be aware of. And, you know, of course, we use a lot of the car data to make that easier and reconcile that automatically. But probably the biggest challenge overall is just building the core financial infrastructure because this is something that, you know, we're literally the only issuer in the U.S.

3:16:42that actually had done this from scratch. And just the amount of time and energy that we dedicated into building this, you know, core financial services infrastructure is pretty tremendous. but it gave us a pretty big advantage when it comes to these enterprise customers. And today we have, maybe three or four years ago, I think we had five or maybe 10 public companies in racks, and now we have 250 and growing. So that's an exciting thing to compound on. That's amazing. Awesome. Well, congrats to the whole team on the deal. Congrats to Art. Yes. Love Art. Chief business officer, correct? CBO. CBO.

3:17:14He was coming in to defend. Keith Raboy last week was talking some smack about the BD people of the world. Does OpenAI have a chief business officer? Because they just got a CRO. They have a CFO and they have two CEOs. I think they're missing a CBO. It's an underrated title. You better lock down Art. He's going to get poached. Art is taken. Art is taken. Don't listen to that. Awesome. Yeah, congrats to the whole team. Congratulations. Congratulations. Great to have you on the show. Yeah. Thanks for having me. I appreciate it. We'll see you soon. Have a good one. Merry Christmas. Goodbye. Thank you.

3:17:48Merry Christmas. The ho-ho-ho really gets me every time. I've been thoroughly enjoying it. I love the idea of country-specific per diems because if I had a German employee, I would want them to know that they are not an American employee and they are second-class citizens in this organization and they will not be dining out on the company dime at a nice five-star restaurant. At least in U.S. dollars. They will be eating sauerkraut in the gutter for$5. That's brutal. The most American American, John Coogan. Yes, yes. Well, speaking of travel to international countries, if you go to Saudi Arabia, you can now booze it up, I guess.

3:18:35You can drink alcohol. Saudi Arabia is loosening alcohol rules, letting non-Muslim foreign residents earning over$13 ,000. You can buy alcohol. So you have to check. You have to do a whole checklist. Daniel's line is great. He says, this is such a galaxy-brained idea. It could only come from a mind totally disconnected from normal material reality. It's hilarious. It is interesting. It's like, hey, if you want to do drugs, if you want to consume alcohol, that's fine. But you have to be putting up numbers. Yes, yes, exactly. You've got to be putting up numbers. Exactly. And so if you're underage.

3:19:11I think cannabis should be like$10 million a year on your W-2. I was about to say. But you know what happens then? So you have everyone in America, you turn 17, 18, 19. You're not 21 yet. You get a fake ID to try and buy some booze. You're going to need a fake W-2 as well. You're going to need to fake your income statement. And you're going to need to do all these crazy things where you have all these circular deals. Bringing your fake W-2 to the corner liquor store and just looking through every line. In Saudi Arabia. So you're in Saudi Arabia. And let's say you're making$1 ,000 a year USD. or you're making 10 ,000 reals, you need to go and set up a circular contract where you're a 17-year-old, somebody else is 17.

3:19:54I pay you 10 ,000, you pay me 10 ,000. And then we get our - Do that a couple of times. Exactly, we do that a couple of times. Then we're earning enough to go buy alcohol. We can go get drunk together. That is crazy. That is truly galaxy-brained. We're out galaxy-braining them. We're going to galaxy-brain all the way to the top. We do have to hop on with Riyadh soon. let's close out with anything else that's on the timeline. Jordy, why don't you take a pass? Let me know if there's anything that we mix up. FluidStack raising apparently$700 million out of$7 billion valuation built on Google-backed leases for new AI data centers, according to Rohan Paul.

3:20:39FluidStack, a real player on the semi-analysis. What's it called? ClusterMax. ClusterMax. Um, and, uh, so not, not surprising here. Interesting that, uh, potentially you've got Leopold, uh, looking to lead the financing. So dipping his toes into privates. Well, the, uh, I mean, they have cluster max over at semi-analysis. They have inference max. Are they going to do looks max? Potentially. They could. It looks like according, you know, looking at some of our officials, it looks like we might be developing a strategic looks maxing reserve. I mean, there are lots of people that would comment on looks maxing and try and rank like the hottest men in tech.

3:21:22I personally would only I would only regard semi analysis opinion as credible on that on that issue specifically. It must come from them. Post here from Buko. He says Hawk Tan acquired VMware, had coffee chats with the employees where he made fun of them. Oh, this is fire to half the business. So a little excerpt here. Earlier last year, Broadcom CEO Hawk Tan decided to host what he likes to call a coffee chat with employees of VMware, the software giant that Broadcom CEO had recently purchased for$84 billion. The company-wide meeting was Tan's effort to introduce himself and answer questions from those employees about life at the semiconductor firm.

3:21:58At the time, VMware's Palo Alto campus sprawled over 18 buildings and 100 acres of delicately pruned gardens, an outdoor amphitheater, and a turtle pond. Employees enjoyed nice HR perks, including childcare, marital counseling, and an annual$1 ,000 well-being allowance for anything from dumbbells to Xboxes. $1 ,000 on Xboxes? That's like one Xbox every three months. That's so many Xboxes. You can actually only spend this on Xboxes. Why do you do that? When Tan opened the discussion up to questions, a VMware employee asked if Broadcom provided such benefits. Tan seemed surprised. Why would I do any of that?

3:22:36I'm not your dad, he replied, according to three people in attendance. We need a trading card. I'm not your dad. Over the next few months, Tan fired about half of VMware's 38 ,000 employees. That is wild. He also stripped down the campus, selling all but five of the buildings. At the remaining offices, Tan had the espresso machines removed. Okay, well, that is a question because they need to stay caffeinated. I got to hold the line here. What are they doing? I mean, I'm okay with, I'm okay removing espresso machines. Do you think they're doing white monsters instead? If they're doing white monsters, then I approve of it.

3:23:08But if you're going to have a decaffeinated workforce, that just makes no sense. That's just not elite performance. He says, the article says, somewhat against the odds, the turtles were allowed to stay. So do you think he's putting espresso in the turtle pond? No, I imagine he had like a massive broadsword and was above the turtle, ready to let it all go. They're like, Hawk, no. And he's like, oh, I can't do it. and he drops the blade and it falls to the ground, clanks him, and he says, oh, you live to fight another day, Turtle. You're too important here. He's got to make in TPUs. Completely different vibe from when Pat Gelsinger was getting a VMware tattoo on his arm, or a temporary tattoo at least, showing his love for the company that he just bought while he was at Intel.

3:23:55Hawk's hand, clearly cut from a different cloth, but clearly also putting up big numbers. Broadcom's been on an absolute run. Alexis Ohanian says been warning y 'all AI everywhere dead internet real is increasingly scarce online and thus increasingly valuable a wired article called AI slop is ruining Reddit for everyone this is my this has always been there's such an incentive for companies and people to just like create narratives on Reddit because AI is training on Reddit and the natural end state is you know it's 99.9 % bots and just a handful of humans so stick to the group chats, I guess.

3:24:34Yeah. In other news, Netflix has launched Best Guess Live, a new Netflix mobile game show where guessing the right answer based on five clues could win you thousands. Dylan Aberscotta on our team... Thousands in this economy? Come on. Not escaping it. Oh, they're willing to spend$80 billion to get$40 billion per Warner Brothers. They should give you some IEP. You should be able to win Porky Pig. I get Foghorn Leghorn. Or some of the masculine films. I want the Steven Seagal rights. I want the rights to hard to kill for sure. Dylan Elberscott on our team says, can I copy your homework? Yeah, just change it up a bit so it doesn't look obvious.

3:25:17You copied, of course. He worked at a competitor to this or the precursor to this. He created game shows. He created game shows with HQ Trivia, which you should know. Rune says, a lot of people think CEOs are on coke or stimulants because of how hyper they look in interviews but i think the unfortunate truth is that natural amount of energy is extremely unequally distributed i was going back yeah i was going back and forth on this one obviously it's a subtweet of carp but um i i i think he's i think he's just like that i think he's just i think he's just high energy and i think that so the the carp clip looks crazy as a single clip, but I actually watched the full interview and there is a flip on it, which is maybe more negative, which is like, that was a question he was squirming on.

3:26:09Like that was an uncomfortable question. And I noticed it on this show all the time. If we ask someone a question and they're like, not that happy that we asked that question, you'll see on their body language, they'll cross their arms. And crossed arms is a, is like a well-established body language for like, I'm closing myself off versus like, if I'm hanging out here with you, I'm just having a good time. If I'm sitting up and yeah. Yeah. And so, and so there's a world where, where, where you're literally, they call it being in the hot seat. They just needed a better camera tracking. So yeah, they call it being in the hot seat.

3:26:41And when you're in the hot seat, you're kind of bouncing around and yes, it can look like you're on a lot of stimulants. Maybe you're on some caffeine. Maybe you're a little wired, but also maybe you're dealing with a serious interviewer, Sorkin, across from you on a stage, there's nothing to help you at DealBook. The crazy thing is you keep seeing people's legs all funny and it's like because you're just in a chair, there's no desk there so there's nothing covering you up. So it's like you're very vulnerable. There's this big wall behind you. Everyone's looking at you and also it's not just a private conversation where it's like you and Sorkin are just hanging out here having a normal conversation.

3:27:18There's a couple people, all of our friends are here, right? Imagine, it's like all of your rivals are in the, all of your rivals. And all your haters are watching online. All your haters are watching online and all the rivals are in the building, in the seats right there, looking at you, answer tough questions. I don't know. It just feels like this is not that crazy of a reaction for a high energy guy who's kind of quirky, who's in that position and is asking, is answering some tough questions. I don't know. But it will continue. Well, Tyler, figure out how to use AI to run an analysis on every guest we've ever had.

3:27:52to see who and just rank their body language from most trustworthy to least trustworthy. Yeah. I did like Gabriel here. He says, he's quote tweeting someone who's like, there is zero chance that these hyper animated CEOs are just naturally that way. Kind of the counter to what I was just laying out as an argument. And Gabe says, cocaine expert here, these CEOs do a lot of cocaine. Feels like a very 2025 moment somehow. And I agree with that. I It is a very weird, weird thing. Well, in less weird headlines, we have some fantastic news from Lulu, who has raised a$40 million dollar bill. Give me the mallet.

3:28:28Give me the mallet. All right, no, you go for it. Too late, too late. Congratulations to Lulu. This photo goes hard. It's a good photo. It's proper paparazzi. She got caught by the paps. The paps got her. This is a throwback photo from when she was working on the Microsoft Blizzard deal. I mean, the fact that she's a board member at Activision is just crazy at the time. $80 billion company. You know, people talk about Lulu like she's some small-time PR advisor for startups, but it's like she was on the board of a$80 billion company. She also intentionally flies under the radar. Yeah, well, we're blowing up here now.

3:29:12Sorry about it. Anyway, at least it's at the end of the show. Lisan Al-Gaib says, the normies are waking up on TikTok. They're saying Chachiput is washed. John says, perplexity is goaded. Hard to say. They're ranking them. They're all over the place. Claude, greater than Gemini. Never be washed. You don't want to be cooked either. Yes, we've learned this. Don't be washed, cooked, or chopped. From our Gen Z adoption. From our Gen Z employee, who is currently both chopped and cooked. He's not here. We hope he gets, not you, the one who's sick, who we hope gets better soon. But we didn't learn any lingo from you, Tyler.

3:29:55Buko says, what people get wrong, OpenAI's Code Red is bullish, not bearish. It's an admission they were overeating, getting beat, and needed to focus. That's what great teams do. Now here's the challenge. Code Reds are really hard. It's easy to say yes, hard to say not right now. All eyes on how they execute Code Red. I'm skeptical overeating seems to be embedded in the culture. Sam obviously likes to dabble in a lot of things Every code red is followed by a Baja blast. That's right. I think the next the next model should be Baja blast go back to the crazy names that make no sense Stop it with the 5.2 tomorrow.

3:30:31Wait wasn't 5.2 just supposed to come out today? That was if you looked at Polymarket, that's what it looked like. It's now moved to I think by December 13th It's like 90%. Yeah, yeah, yeah. Anyway. But also, this is a separate topic, but I need to put you in the truth zone. There actually are spots on the moon that are perpetually in sunlight. Oh, really? You figured it out. On the rims of certain craters. Oh. The inside of them are always in darkness. The rims are sometimes in darkness. The rim on the crater is always in the sun. Yeah. So that's where we need the data center. Okay. Well, how much crater rim is there?

3:31:04Is there enough crater rim to actually build a lot of data centers? If there's not, you could probably nuke the moon. Make more craters. Yeah. Yeah. I don't know. I still have trouble visualizing this. Doing like the celestial body physics math in your head is very, very difficult. In other news, this person on X, Kinu Nazari, says they got their hands on some North Korean cigarettes. They did? They got them in a DPRK state-owned restaurant from a waitress who was from Pyongyang. basically impossible to find any background info on them online they taste and look 100 percent what is that i have no idea uh but i like the packaging the packaging goes pretty hard it feels uh does but i can't support it because i don't support north korea uh and this is a good place to end the show sag harbor capital says if this niche scandinavian brand doesn't ship my overcoat in the next 24 hours i'm gonna blow up the nordstrom 2 pipeline

3:32:10he tags the brand too he does wear oh wait bergenberg i need an overcoat i was trying to just go around to york last week just my suit and i and i was kind of cocky going into it i was like oh john nice nice coat you're you're gonna be outside for two seconds 20 seconds yeah you're In succession, Kendall Roy doesn't need a coat because he's so rich. I didn't say that. I didn't say that. I think Dylan said that. Dylan said that. But I was like, yeah, we're not that rich yet. Whatever's going on in succession. We went for a half-mile walk, and I was devastated. We did. But it's nice when you're in a coat.

3:32:48It's fun to walk around New York City. Also, it wasn't that cold, but you were getting destroyed. We got to pull up last post of the day. uh skooks says the concept of getting sent to rehab by the taliban because you and your cousins wanted to larp as the peaky blinders in afghanistan and uh what happened here that says reportedly four people in the herat province were detained and sent to rehab by the taliban's virtue ministry for imitating characters from british crime series peaky blinders and parading around the city resembling the vintage gangsters an act seen as violating local values this it seems I have no idea how to process if this is real at all.

3:33:27Hope says rare Taliban L. This is crazy. I mean, they look fantastic. They do look fantastic. I feel like this would encourage, potentially they could open up a tourism market for people, kind of like the Disneyland for Peaky Blinders enthusiasts. Such a weird phenomenon to happen here. Also, like, what does being sent to rehab look like in Afghanistan? They're just like. No more Peaky. Is it like Alcoholics Anonymous? You're like, first step, I am powerless to Peaky Blinders. I am a Peaky Blinders impersonator. You have to come, you know, Alcoholics Anonymous, you have to say, I'm an alcoholic.

3:34:06You know, I'm a Peaky Blinders impersonator, and I am powerless. And you need to say the Lord's Prayer, something like that. Anyway. Anyways, thank you for being with us today, folks. we love you dearly and we will see you tomorrow we hope you have a fantastic evening goodbye cheers