TBPN
SpaceX Might Be the Greatest Company of All Time
Ask the tutor about this video
Sign in with a verified email and ask the tutor anything about this video.
TL;DR
Research summary
▶ SpaceX’s near-term drivers
The discussion opens with praise for the SpaceX IPO execution and a question about what investors will focus on next. Baker says quarterly results matter, but public markets have shown more tolerance for investment and longer time horizons than some venture investors assume, citing Tesla and Amazon’s infrastructure investment cycles. He identifies two nearer-term variables: how quickly SpaceX can bring terrestrial compute online and Cursor’s prospects. He says SpaceX appears to monetize gigawatts at a higher rate than peers; the discussion cites an Altimeter estimate of roughly two to three times average neocloud pricing and a $50 billion per gigawatt figure for the Google deal. Baker says bringing two to four gigawatts online in the next year could mean substantial revenue, although prices could rise or fall. He regards space exploration and asteroid mining as exciting possibilities, but says more tangible near-term drivers deserve attention.
▶ Compute allocation and hyperscaler strategy
The hosts ask about tension between selling cloud compute and reserving it for SpaceX’s own applications and models. Baker suggests SpaceX’s willingness to make external compute commitments may reflect confidence in bringing gigawatts online, pointing to reported orders for 20% of Rubin chips and expected Groq LPU integration in the next six to nine months. The speakers note that compute may be needed before all related revenue arrives, and discuss possible flexibility in the Anthropic agreement and the Google deal. Baker says a token-factory business could be enough for the next five to ten years; if broader cloud services become necessary, he expects SpaceX could build them. On Meta, he sees substantial data-center assets but says the market appears skeptical that the company can monetize them. He considers that skepticism warranted given the limited public evidence of a successful AI product, while noting Meta could apply GPUs to improve its advertising business.
▶ AI competition, Meta, and retail investors
Baker says the behavior of major companies suggests they believe AI competition may be entering an endgame. He links OpenAI’s reported Codex price cuts to the value of coding tokens and the pursuit of a recursive self-improvement loop. He describes Meta’s Muse Spark as an upside surprise that made him think there was a chance of meaningful progress, and says Mark Zuckerberg has shown an ability to pivot. He recalls Meta shifting quickly from saying it would not monetize GPUs externally to emphasizing operating-expense focus, and says the company’s stated position could change. Baker argues that stock performance also matters to AI companies because falling share prices can hurt employee compensation and retention. On retail investors, he rejects treating the label as inherently pejorative and says retail favorites have performed strongly, in his view outperforming many professional managers. For SpaceX, he speculates that some IPO retail allocations may have gone to flippers, while saying his understanding was that institutional buying was greater after the IPO. He also says 10 of 10,000 SpaceX employees bought in the IPO, and argues that employees and existing investors have had prior opportunities to sell. He expects supply and demand dynamics to be interesting and says retail investors have been an unusually powerful market force over the prior three and a half years, while distinguishing the current period from the 2000 bubble.
▶ What counts as being in the token path
The hosts ask whether Cloudflare and other content delivery networks (CDNs) count as part of the token path because they can deliver tokens at the edge. Baker says CDNs can provide low-latency delivery to valuable users and command a premium for speed, but estimates that they currently deliver less than 1% of tokens consumed globally, possibly less than 10 basis points, because most processing happens internally before users receive results. He says CDNs have part of their business in the token path today and a path to a larger role. The discussion contrasts low-latency voice interactions with coding tasks that can run for longer in a central data center, and mentions proposals for GPUs at the network edge or even near homes. The speakers compare local compute with local energy storage, then turn to upstream bottleneck companies. Baker says the bottleneck trade may be nearing its end and that attention is shifting toward businesses with enduring franchise value.
▶ Orbital compute and sovereign AI
Asked whether land is a bottleneck for compute, Baker says orbital compute could become economically compelling if Starship becomes reusable. His estimate compares a terrestrial gigawatt costing $60 billion, including $25 billion for power and cooling and $35 billion for IT equipment, with an orbital gigawatt: assuming $5 billion launch cost for one reusable Starship, he estimates $30 billion total. He calls the terrestrial power-and-cooling cost reasonably inflationary, but is unsure whether land itself belongs in the token path. On sovereign AI, Baker expects every country to want a strategy at least for national defense, but says he does not see most countries building frontier AI. He expects countries to use a leading provider or open-source model, adapt it through reinforcement learning and supervised fine-tuning for local language, culture, and values, and run it in domestic data centers for defense, intelligence, or policing uses. He expects a substantial buildout for those applications.
▶ China’s models and distillation
Baker attributes a widening gap between Chinese open-source and American closed-source models partly to what he calls China’s mistake in not taking the opportunity to buy advanced chips, and to an asserted belief that domestic chips are sufficient. He says Chinese labs are particularly effective at distillation. Citing something he says he was told, he gives 160,000 reasoning traces from O1 and O3 as sufficient to produce the original DeepSeek, and describes industrial-scale use of APIs and large numbers of endpoints for distillation. He argues that this advantage could diminish if frontier models are no longer released, and views Mythos as a sign of that possibility. The hosts add that labs have become more aware of distillation risks and are better at protecting reasoning traces.
▶ SpaceX as a rare, potentially iconic company
Asked whether he may never see another company like SpaceX, Baker says he is not worried but calls his experience with it special. He says he previously owned 15% of Nvidia when it had a market capitalization below $2 billion and 10% of Tesla at a market capitalization below $2 billion. He says SpaceX may become one of the most important and iconic companies ever, while acknowledging the engineering required for ambitions such as cities on the Moon and Mars. He defends the company’s perceived benefits, citing decarbonization, low-cost internet access for schools and hospitals in low-income countries, assistance for blind people, and wealth created for blue-collar workers. These are Baker’s arguments and opinions. He recommends seeing a rocket launch, describing the experience as visceral and inspirational. He is skeptical that another company will provide a similar experience and says that if SpaceX achieves its larger ambitions, it could become the most important company of his lifetime or possibly of all time.
▶ Small-cap investing risks and close
The final discussion turns to opportunities below $2 billion in market value. The hosts mention possible AI spillovers into biology and materials science, and Baker says there are many venture startups at that scale, though he also says little was happening in public small-cap markets before AI and the bottleneck trade. He expresses concern that people with large social-media followings may promote stocks in companies with market capitalizations as small as $35 million, potentially sending prices up sharply; he says there is a lot of questionable activity, while acknowledging that a person discussed is smart. The speakers compare the situation with the dot-com period, when market excesses were followed by bankruptcies and civil penalties, alongside the emergence of enduring companies. They note that anonymous accounts and limited accountability make the current environment different from regulated sell-side research, and describe it as uncharted territory. They close by advising listeners to stay safe and do their own research, then agree to meet again with the wider team.
Full transcript
[00:30] execution from start to finish. Was that how you interpreted it? >> I thought Goldman Sachs and Morgan Stanley did a very good job. >> Yeah. Um what do you think the market is looking for SpaceX next? Is it all eyes on the first earnings report? Like where do we go from here? Because there's so much of the SpaceX story preipo that was a decade out, 5 years out, two years out. But is the market going to be processing quarter by quarter plays like a like many other companies? >> Well, I think the market is always, you
[01:01] know, quarters matter. >> Yeah. >> Um, you know, the runners on X, you know, it's like they're these reply people on X for any topic or experts. >> Yeah. >> But, um, somebody once told me a marathon is 26 onem runs. Yeah. >> And so, you know, and then all the runners were like, "No, that's not true at all." Running, you know, obviously running a mile is different than running a marathon. But, you know, every quarter matters. Uh but at the same time I do think if you look at how um the market has has looked at Tesla
[01:31] over the last five six years you know Amazon during kind of the days of you know when they were building out AWS and then um you know kind of their their retail distribution infrastructure when they'd go through these investment cycles. I do think the public market has a much greater tolerance um for investment and a much longer time horizon than a lot of people in the venture ecosystem give it credit for. By the way, this is Foxy in the back. Um he's our he's um he's one of the
[02:03] analysts here at Trades. >> I was just listening to him on the Brad Gersonner podcast. I think he chimed in on that, right? You brought him in. There we go. >> Yes, he did. But he did the work on SpaceX. So, I thought you guys might want to fantastic about that. Um, but I do think that I'm not sure the story is as far out as you made it seem to be. >> I think there are two variables that are going to matter a lot over the next year. >> Mhm. >> The first is just how quickly can they bring on terrestrial compute. >> Yeah.
[02:33] >> It does seem like they monetize gigawatts at a higher rate than anyone else. >> Yeah. >> And we know from Jensen that they bring on data centers faster. >> Yeah. >> Than anyone else per Jensen's words. Yeah. >> Um and everyone in the ecosystem is really incented to get um land and power in the hands of people who can energize it because everybody starts making more money when the GPUs are energized. Um so you know if they can what what are what
[03:03] what was the altimeter figure for how what they were monetizing gigawatts at Foxy? Do you remember for which deal? Well, I think it was like two to three, wasn't it? Like at least two times higher than like average Neocloud pricing and that's partly in due to scale and customer quality and things like that, but it was meaningful. >> Yeah. So, they're doing 50 billion a gigawatt on the Google deal. >> Wow. >> So, how quickly they can add gigawatts really matters. >> Yeah. >> Right. If you can energize two, three, four gigawatts >> y >> in the next year,
[03:34] >> that's a lot of revenue. Now you know that prices may go up, they may go down, but that's really going to matter. >> So all eyes on Colossus 3, 4, 5, 6, 7, 8, 9, 10, etc. At least in the short. >> Yeah. Or macro hard, macro harder, macro >> hardest >> to that name. Yeah. Macro hardest. >> Yeah. >> Um >> and then I think cursor is another big variable. >> Yeah. Um, you know, we we talked about it with Brad about how composer 2.5 after three weeks of, you know, RL and
[04:06] supervised fine-tuning Colossus 2 is kind of paro dominant. So, what's going to happen when it's applied to a bigger, better base model? Um, I do think cursor being in, you know, half the Fortune 500 >> is interesting. So, I think those two things, you know, while I mean, I'm super excited for going to Mars and asteroid mining and >> and you know, a city and mass drivers on the moon, like all that stuff is awesome and I think there's a, you know, decent
[04:37] chance it happens in my lifetime. I just think there is there's much more tangible near-term drivers here. >> Yeah. How do you think about the tension of the cloud business versus their own application and and model business? Uh I think a lot of people were surprised when the initial anthropic deal happened purely because uh knowing Michael Truel, he would have been very excited. >> That's my compute. >> Yeah. Yeah. Get access only child around here.
[05:07] >> And it and it feels like uh you can imagine SpaceX, you know, bringing on a lot more compute very quickly. Uh but at the same time, every lab has had to go through this tension of like how much how much compute do we allocate to training versus inference. >> A way to interpret that is that maybe the team at SpaceX is pretty confident in their ability to bring gigawatts online. >> Yeah, that makes sense. >> And the altimeter figure that they've ordered 20% of the Reubins and the Reubin is an epic chip. Like Blackwell,
[05:40] it was really hard to get it online. Ruben is kind of a more drop in replacement. >> It's a really really good chip. You'll have the um you know the Grock LPUs integrated at some point in the next six nine months. >> Um >> so one way to interpret that is they're pretty confident that they can bring data centers online pretty quickly. Um, but I also I I don't remember all all of the details, but I do think there there was an out in the anthropic agreement that
[06:12] if they >> Yeah. And the and the Google and the Google D could be short-term. Well, and that's that's what I just meant. It's like it is there is like a tension there where if you you know there's a moment where maybe you want the compute, but you don't necessarily have all of the revenue yet and so you could see Anyways, I'm I'm very confident they can figure it out. >> Yeah. on on on the topic of like Elon Web Services, is there a world where uh SpaceX winds up building out more of what looks like AWS? Like we've seen between OpenAI and Anthropic, there's this tension of like you got to be on
[06:42] AWS because it or you know the ecosystem or enterprises just want to be on AWS. They won't even go over to Azure or Google Cloud for whatever. And you could imagine like it might is it enough to just be a token factory or do you need to have you know a database and an in-memory storage and you know cloud storage and cold storage and all the things that ABS provides. >> My thought is that being a token factory is more than enough for the next 5 to 10 years man. We are
[07:12] >> yeah we need more token factories. Um but we'll we'll see and if they need that stuff like they will build it. >> Yeah. How are you thinking about the other hyperscalers, the other big tech companies? Uh we we were sort of puzzling over uh Meta's strategy at this point. Tons of resources, cash flow, data centers. They have experience there. Maybe they're not as fast as as SpaceX, but there's a lot there, but it's a completely different motion to go into token factory. Um and yet there's
[07:42] still a lot of demand. So, how do you think all of that plays out? >> Well, I think there's many ways to go into the token factory business. You could if you're if you're meta, >> yeah, >> you could just say to someone like Fireworks, um, >> hey, >> here are 100,000 GPUs. We'll we'll take a revenue split and we'll see what happens. Y >> um I am increasingly looking at EV to net PP&E as a valuation metric just with the idea being installed atoms on earth I think are going to appreciate
[08:14] in value and this is a little bit that halo trade that um I think maybe Goldman Sachs came up with you know high asset value low obsolescence >> but Meta you know Meta's EV to net PP& multiple is in an interesting place it just suggests that the market has an immense amount skepticism about their ability to monetize their own asset base. >> And and I think that's I think that's warranted given that all we've really got out of management is this concept of
[08:45] personal super intelligence. >> And it would be different if Meta AI was like at the top of the app store and everyone you were talking to was like, "Yeah, you know, I'm using Meta AI >> where there even there was like a killer feature within Instagram. It's like, oh, people are shopping in Instagram now." And I think no one no one doubts that like once Zuck has an AI product with a billion dollars of revenue, he can take it to to tens of billions, >> but it's unclear. I feel like, you know, going into enterprise, like, you know, what is this API business going to going to look like? We know they can use the
[09:15] GPUs just on making the ads product better at the end of the day. And um >> but there's a lot of questions. >> Look, I think a few things. I do think it if I look at the way different players are behaving, it feels like everyone thinks we're entering the end game. Everybody, you know, loves to use these chess analogies and it looks like the endgame maybe here sooner than we think. I think that's why OpenAI is cutting codeex pricing >> simply because coding, you know, coding
[09:46] tokens are so valuable >> and if you don't have enough of them, you may not be able to get into this RSI loop that that everyone is, you know, >> pushing for. >> Um, and I I think if you're Zuckerberg, you feel that acutely. I thought muse, what did it? >> Muse spark. It it was to me it was an upside surprise and it made me feel like you know what's the uh what's what's what's from Tallaladega Nights or what
[10:17] is it? So you're saying there's a chance. >> Oh yeah. >> Like yeah so you're telling me there's a chance. Um >> I mean the other quote from Tallaladega Nights is if you're not your f if you're not first you're last. >> And and you know that's how there are so many good quotes from Talladega Nights. >> Sold the windshield. >> Yeah. 100%. But I think if um I'm pretty confident Mark Zuckerberg has proven to be a good entrepreneur over time and has been willing to pivot, you
[10:48] know, and it's just it's so funny, you know, right now Facebook is, you know, Meta has tell investors there's like no chance that we're going to, you know, you know, monetize GPUs externally. Well, like 10 days before they announced all those layoffs. Um, which I think was at the end of 22, if not I don't remember when they started to announce the layoffs. It was short shortly after Brad, maybe at two months after Brad wrote that letter to Mark where he said, "You're the king. You can do whatever you want, but this what I might
[11:19] respectfully encourage you to do." Um, you know, they went from saying, "We're not going to do that. we're gonna keep investing to, you know, super focused on OPEX in a couple of days, man. So, it's like people they meet with Meta and they hear, "Oh, you know, we're not um we're not going to monetize our GPUs externally." Well, I don't know, check back in a few hours. >> Yeah. Yeah. No, it's like everyone it comes down to like everyone has a price. You look at you look at Elon and Anthropic relationship four months ago.
[11:51] uh and then you look at kind of kind of as the opportunity unfolded >> it made sense for them to partner. I wanted to ask how >> on the meta thing I think there is one point that's important what ultimately drives behavior for all these big cap companies like getting to AI I think that does feel um existential for all of them but the stock price also really matters sure >> because if you don't if the stock goes down you know you gave people a grant at this price stock goes down you know your
[12:22] engineers are unhappy and it's tougher to get to where you want to go if you don't have you know these you know Oh, forget the 10x engineer. People are talking about the 10,000x engineer. >> You know, you're not going to get there if you don't keep those 10,000x engineers. >> Yeah. >> So, just I think >> like we'll see we'll see with Meta. >> Yeah. >> Yeah. Uh how did this IPO uh update your your general kind of framework for the role that retail is playing in capital markets? Uh a lot of uh it was very
[12:54] funny last week to see people like uh say like oh that the IPO is only 2x overs subscribed by retail all these other you know three billion IPOs were we're 10x over subscribed and it's like no we're we're talking about you know a very you know order of magnitude difference here but it still feels like retail is going to >> play a very very important role over the next 12 months maybe more so than ever given that the the the hyperscalers are are now running um you know cash flow
[13:26] negative. You have the conflict in the Middle East, right? There's rebuilding efforts that needs to needs to happen. And so I feel like retail this is this is kind of um retail's moment now maybe more than ever. >> Yeah. Well, so a few things. One, people say retail in a porative way, and I would just say stupid is stupid does. And like you know we track you know there's all these indices of retail favorites and man they're up a lot and >> 25 and they're up a lot in 24
[13:58] >> you know so kind of stupid as a stupid does. So I'm not um you know like I think retail is probably outperforming the overwhelming majority of professional money managers whether they're you know private equity managers venture >> you know public equity managers whatever. So like retail, I don't think it's pjorative. I do actually think in the case of um >> SpaceX, I'm not sure, you know, you process so much information. I'm not sure where this comes from, >> but I actually think after the IPO
[14:29] um you know, for whatever reason, it some of that retail stock may have may have gotten allocated to people who were flipping it because my understanding is there was more institutional buying. I don't know where this after the IPO was retail buying. Now, I'm sure you know retail is very sensitive to momentum. So maybe what's happening today changes that. But I think another really underappreciated thing about the SpaceX IPO is that 10 over 10,000 SpaceX employees bought on the IPO and that's
[15:01] just you know everybody does all these lockup analyses and oh how much stock is going to come come to market and you know for sure there's all these triple layer SPVS or whatever you know or maybe there are there are people speculate there are um and I'm sure some of that will get unwound >> but the reality reality is is that SpaceX employees own a lot of this. And if you're an employee or even if you're an investor who has an S an SPV or you're directly on the cap table, you've
[15:31] had a chance to sell every six months for the last 10 years, >> if you wanted to sell, you could have sold in, you know, whatever when they did that offering in in the second half of last year. So I think the supply demand dynamics here are going to be very interesting. But with retail in in general um yeah man they're they've been a more powerful force in the market over the last three and a
[16:01] half years than at any other time in my career including the year 2000 bubble which I think this is nothing like by the way. Um but yeah, they they matter and stupid is and stupid does. >> Uh I'm interested to go a click deeper on the idea of like value will acrue to companies that are in the token path. I think that phrase is uh extremely powerful right now. But I want to sort of expand it and understand the fringes of the token path because uh we talked
[16:32] to Matthew Prince at Cloudflare stocks at all-time highs. It's delivering tokens at the edge and obviously there's an immense amount of AI agents that are interfacing with different websites. They need to be distributed. So there's a lot of value there. But is Cloudflare in the token path if they're the one finally sending you the packet of tokens at the edge >> or does that not count? How does that how do you think about that? >> Every CDN has has this business. You
[17:03] know, they signed >> Yeah. what was it a$ 1.8 billion dollar with anthropic and if you have all these points of presence yeah >> and you can deliver really low latency tokens to highv value users >> I think what's interesting for for cloud for all these CDNs >> is they're they're getting a massive premium for that latency >> you know it turns out you know one of the lessons from from cerebrus is people are willing to pay for speed >> and you see that in the prices these guys are commanding but I do think I mean my gut is in terms of tokens
[17:34] consumed um on planet earth. These CDNs are delivering less than 1% of them and I mean maybe less than 10 basis points. >> They're all most of the time they're happening internally and then you just get the result. >> Um >> would I say that these guys are in the token path? I mean I would say they have a path towards being in the token path. Yeah. >> And they have part of their they have part of their business that's in it today. >> Yeah. And they're trying and like everyone they're trying to move it more and more there.
[18:04] >> Yeah. Yeah. I mean uh Matthew was talking about uh h colllocating Nvidia servers on the edge specifically for the delivery of voice models because you want those to be very low latency. Maybe you want your coding model to go off and cook in Virginia for 20 minutes or an hour and then come back to you with the final result. But uh if you're trying to actually talk to a model, you probably want it as fast as possible. What about on the opposite end? Wasn't there that startup that Nvidia just invested in that like will >> Oh yeah. on your house >> money >> that put a GPU you know four a four GPU
[18:36] server on the outside of your house. >> I mean >> GPUs are going everywhere going in your bed eventually. >> I mean it's full circle too. Wasn't Jensen talking about stuff like that back during like the original crypto cycle being like everyone's going to be you know mining. >> Yeah. And I mean the the Tesla Power Wall is like a version of that for storing energy. It doesn't store compute, but it stores energy locally on the edge, off the grid, and uh there's a lot of powerful things that get unlocked by that. What about land? Is land in the
[19:06] token path at some point on the extreme other end? Uh because people are obviously searching for go going further and further upstream. They found a a toilet company that makes a particular material that goes into semiconductor supply chain. People are hunting for like the final the final ingredient. Maybe it's just sand that turns into silicon. Yeah, I do think the bottleneck bros as they are called are this like bottleneck trade is kind of nearing its end. >> Okay. >> And you know I did um
[19:37] >> you know the Wall Street Journal had the story about this Japanese company that makes um I forget what they make >> Yeah. >> Aimoto. Um, and you know, everybody's in this because they thought they were going to raise prices and then they just said, "We're actually not going to raise prices. This is the wrong thing for us." And I almost posted yesterday on X, welcome to Japan to all of the uh bottleneck people. But, you know, everybody's, you know, having Claw and Run, what's the next bottleneck?
[20:08] >> Yeah, >> that was that was the game for the last year. The next game is what has enduring franchise value. >> Yeah. kind of on the other side of these bottlenecks whenever whenever that is >> as far as land is land a bottleneck man I do think the math for orbital compute to me it once they can reuse Starship >> I think the math for orbital compute becomes pretty compelling you know it's just 60 billion to bring on a gig terrestrially 25 is power and cooling
[20:38] you don't need that in space >> and so the right comp for that 35 billion of kind of you know IT equipment GPUs CPUs switches um you know memory storage um is is that 25 billion versus the cost of launch and one starship is reusable I think the cost of launch is 5 billion so 20 that means you can put a gig into space for 30 billion and a gig on earth is 60 billion and that 25 the power and
[21:08] cooling >> feels reasonably inflationary to me >> sure >> um so I don't know that I would say and is in the tokens token path. >> Maybe beachfront property, but that's about it. >> Yeah, beachfront property is beachfront property, airplanes. >> Yeah. >> Really nice cars. >> Yep. Firmly in the token path. >> All that stuff is in the token path. >> I gota uh how do you think the events of of last week with Anthropic and DC
[21:38] update uh different countries on their own sovereign AI play? Because there's been there's been attempts It doesn't feel like, you know, many of the attempts have, you know, you can have a powered shell and you can have the best GPUs, but if you don't have the talent, it's really, it's going to be really hard to get to the frontier. And now, if you don't have massive usage, it'll also be really hard to get to the frontier. And so, do you think the ship has sailed? >> No. Well, look, look, I think every country is going to want to have their own sovereign AI strategy at a minimum
[22:08] for national defense. I think where that ends up is um I saw today that the EU had restricted exports of mrols. >> I think that's a joke. I think that's a meme. So people have gone so far on the MRL leashhat thing that that they're saying it's like mythos is distilled on it and it's breaking all the benchmarks. I think it's a good model, but I don't know that it's quite there yet. >> Yeah. >> But you can imagine that we get there
[22:38] for sure. >> Yeah. So I think what sovereign AI will look like for essentially all countries really other than the United States and and China, although I think China is going to fall further and further behind >> is just some sort of you do you use um you know you use one of these providers to do some reinforcement learning on your language, your culture, your values. >> Um you have a system prompt. um you do some supervised fine-tuning and you run that on whatever the best open source
[23:09] model is. I think that's where sovereign AI ends up. And then you run it in your own data centers so you feel like you know it's safe and whatever defense questions you're asking it or whatever forget questions whatever defense and intelligence um and maybe um you know policing activities your sovereign AI agents are doing you know 24 hours a day. You feel like they're safe. >> Yeah. So I think that's where sovereign AI ends up and I think you know there'll
[23:39] still be a big buildout for that but man sovereign AI at the frontier I don't see it. >> Yeah. What do you think the drivers are of the widening gap between Chinese open source models and American closed source models is? >> China's made a terrible mistake not you know taking it feels like the administration has been willing to let them buy H200s or P30s. >> Oh yeah. And I think they're on this like, you know, they have this crazy belief that, oh, you know, our own internal chips are good enough. They're
[24:09] not. >> Yeah. >> And then, you know, I think what's making them think that is that Chinese labs are very, very good at distillation. Somebody told me it only took 160,000 reasoning traces >> from 01 and 03 to get the original deepseek. You know, they're very clever at industrial scale distillation, running it through multiple APIs, >> pinging each API from, you know, we've all seen those iPhone farms in China. You know, they've got they've got the
[24:39] same thing for distillation and they've got, you know, 100,000 endpoints, distilling these models across every API available. They've gotten really really good at that, but man, all that goes away if people stop releasing these models at the frontier. And I think mythos is a side of things to come there. >> Yeah. Yeah. It seems like they're getting much better at locking down the reasoning traces, locking down the the espionage. It feels like uh back during the 01 days, the labs weren't even aware
[25:09] that distillation was so possible that that was uh that that was something that was such a such a potential threat. For sure. >> George, please. >> Is any part of you worried that you'll never see another another company like SpaceX? Um, >> worried is the wrong word. >> Not not worried, not worried, but but like you have to appreciate you have to appreciate that a moment like that you it's possible that you have an opportunity to invest in a company like that one time in your career. And thank
[25:40] God you did it because there's plenty of people that that had the opportunity and didn't. And uh >> well well I did own 15% of Nvidia when it was a sub2 billion market cap but I own 10% of Tesla when it was a sub$2 billion market cap. Um but um but SpaceX certainly it's been a special experience for me >> and I think there is a chance that it is you know one of the most important
[26:10] iconic companies of all time. I mean it already is. Uh maybe there's a chance it's the most. Um and I think in general it's just like I I kind of want to take a step back and like all this trillionaire hate from the left. It's like the you know there was a great post from my friend um Kevin Mafy saying like oh this is a bond villain who's decarbonizing the world. Last time I checked, people on the left
[26:40] >> were like the environment >> connecting u you know poor people in low-income countries, schools and hospitals to the internet at very low cost and he's helping blind people see today and you know interact with the world. That's a villain. Um and then just like there's so many parts of the SpaceX story like all of the you know bluecollar workers who've gotten super wealthy. um if you're opposed to data centers on Earth because the
[27:10] environment, well, SpaceX is your is your solution. So, um you know, there was another great post that just said, you know, the left won and they don't even appreciate it. The world's first first trillionaire, >> you know, has done more to kind of solve the environment than like everyone else on Earth combined. >> Y >> I saw there was another good one. It's like, yeah, Elon could solve world hunger by, you know, market selling every, you know, uh, piece of stock that he has and send the world into a global
[27:41] economic downturn, losing, you know, millions of millions of jobs in the process. >> Yeah. That uh, somebody said that Elon that the World Bank could solve hunger with five billion. And he and he said, "If you prove it to me, I'll do it." >> Yeah, of course. >> And turns out you can't solve world hunger for $5 billion. >> Yeah. Um, but yeah, SpaceX is certainly I would say, you know, SpaceX, Nvidia, Tesla, they've all been very special to me in in kind of a different way,
[28:11] but man, there is nothing like a rocket launch. And so, I just don't think as a visceral experience, I highly recommend everyone listening, you guys, if you have kids, take your kids to a lodge. You don't want to go with SpaceX. Everybody's always like, "Oh, get me the special tickets." You just literally you could go to a beach at Bokhica. Yeah. >> And be where the special SpaceX viewing area is. You could literally be closer to the lodge. Yeah. >> You could there's hotels near Vandenberg. Take your kids >> see a launch. It's super inspirational
[28:41] that humans are capable of this and it's much more visceral than you realize. >> You know, the sound, the blast of hot air, a lot of people cry. So go see a lodge. there will be a ne never I don't think a I'm pretty skeptical that there will be another um company that delivers a visceral experience like that and then listen if SpaceX puts a city on on the moon and then Mars and
[29:11] enables you know becomes the British East Indian company the solar system I think this will and those there's a lot of hard engineering that needs to happen to go into all that this will probably be the most important company um of certainly my lifetime of maybe maybe of all time. Um and so from that sense, I don't think there will ever be anything else like it. And yeah, I guess that is a little bit sad, but you know what? There's so much awesome stuff happening in the world that it's it's it's tough for me to get too down, man.
[29:41] >> I love it. >> Yeah. Hard hard to hard to be hard to be >> I have I have one last question. We have uh um you mentioned owning Nvidia sub two billion. What is the health of the sub2 billion uh market right now? Is there is there anything that you're excited about? >> Is there anything that the bottleneck bros haven't ran up? >> Haven't run up to pros have taken everything under two billion above. >> Maybe not even maybe not even uh like uh like public companies but private areas. I mean we're starting to see the knock-on effects of AI in bio and uh
[30:14] material science and like Neurolink was sub two billion for a long time, right? And so there are more like not in theme but accelerated by. So I don't know there's also just like >> SAS company >> there's loads of incredible venture companies you know startups under two billion. Yeah. >> Um like loads of them. Yeah. Um you know I I I I talked to them you know several times a week minimum. >> Yeah. in public, you know, until the
[30:44] bottleneck bros came along, until AI ca came along, there was not a lot happening in public small cap land. And by the way, on this like all these small caps, these bottlenecks, this AI, >> it does feel a little weird to me that you could have like a large following on social media, post about, you know, buy buy stock at something. >> Yeah. and then post about, you know, outline your thesis in a $35 million
[31:16] market cap company and send it up like a thousand%. Um, and there's a lot of that going on. And by the way, this not this is not like a serenity comment. I think that guy's that person >> is smart, >> but I mean there's there's there's there's a lot of shady stuff happening it feels like to me. >> Sure. >> Well, yeah. And it you look back at uh you know you were a young uh young analyst at at Fidelity right during the the doc you know the end of the the.com cycle and there was a lot of just like
[31:47] euphoria stuff happening and in the years that followed there was consequences right there there were people ultimately had to to face um >> sometimes civil penalties all sorts of different things companies went bankrupt there were all sorts of different uh you know perturbations in the market. And of course, we got a co we got a bunch of really enduring generational companies through that period, too. >> Although a lot of these accounts are anonymous and I'm sure they're logged into social media with a madeup email.
[32:17] >> Yeah. >> You know, gone through a VPN. And so >> I don't I don't know. We'll see. >> It is different than a research than a cellside research report that maybe goes a little bit too far, but there's still a whole compliance department around it. Uh we are in Yeah. uncharted territory. So, uh, everyone needs to stay safe out there and do your own research, I suppose. >> Yeah. >> Anyway, >> well, guys, man, this was super fun. >> This was great. Let's do it again soon. >> Come on more. >> Thank you so much. Anytime. Let's get the whole team on, too.
[32:47] >> Yeah. You know what? That's actually fun. >> Let's do that with the whole team. >> That' be amazing. >> Awesome. >> I would love it. Yeah. Yeah. You got a conference room with a with a nice wide angle camera. You can just have the whole crew there. That's great. >> Partner into a partner meeting. Exactly. Awesome. Great to hang with you. >> Have a great rest of your day. Thank you so much for coming on.