China Open-Source, Compute Arms Race, Reordering Global Trade | BG2 w/ Bill Gurley and Brad Gerstner

0:00 15% on all goods coming from Europe, 0% on US goods going to Europe. So an opening up of Europe markets, paying us 15% and on top of that getting commitments like $750 billion, almost a trillion dollars of energy purchases from the US. Or look at Japan, which they announced last week, another huge market. Again, similar. They're going to pay tariffs to the United States. No tariffs imposed on the United States. and they're going to invest $550 billion dollars into the US in a way the president gets to direct. So I would love I would I just think we need to give the president credit where credit is due. Everybody said this was going to lead to retaliation to trade wars was going to be disastrous for the US and all we've seen so far is deals deals deals deals. And I have to say, if this was the CEO of one of our companies, right, let's say we had a board meeting at the start of the year and he outlined these plans and we said, "Hey, we're really nervous about this. You know, this is a high-risisk, highreward strategy. It's either going to backfire and we're going to fire you or it's going to work really well and we give you a bonus." If we're measuring them halfway through the year, I would say that he's in line for a bonus.

1:21 [Music] [Applause] [Music] All right, the summer pods back in action. Uh, good to see you guys. We have our good friend Sunonny Madra, the COO of Grock, joining from I don't know, Sunny, it looks like some fancy hotel in Saudi Arabia in the middle of the night. Good to see you. >> You picked it up. You picked it up. You have a good You have a good eye for the Middle East. And Bill, Bill, you got off your boat catching blue fin tuna. You're uh you look like you're in some office somewhere. So good to see you.

1:57 >> I am uh borrowing Mitch Lasky's incredible podcast setup in in our Woodside office with the uh Oh, nice high highde SLR camera. >> Nice. You're looking you're looking good. Sunny, of course, our good buddy Sunny is the COO of Grock. Um I don't know, Sunny, probably a few hundred million of revenues, maybe doubling year-over-year. I just saw something. You're rumored to be raising 600 million at a $6 billion valuation. Maybe that has something to do with you being over in Saudi Arabia. Um, and of course, you're hosting all the uh open-source models in in your inference clouds around the world. Is that about right.

2:35 >> Yeah, you got you got it right. You touched on all the key points. You know, we don't comment on speculation, but you touched on some good points. >> Well, it's great to see you in DC last week. Um, you know, our good friend David Saxs is really on a heater. First, it was the crypto summit a couple weeks ago. Of course, the Genius Act got passed, which really teed up these stable coins and and and now the Clarity Act around market structures making its way through Congress. And then of course last week was the AI summit where the president laid out um a multi-prong strategic plan for American AI that both extends the government's investment and and leadership um but also accelerates the distribution of the American AI stack around the world. Both of those things, both the crypto and the AI summit that he put together, I thought are key contributions really to the next generation of American technology leadership around the world. It's amazing to see that much progress in six months. Congrats to Sachs >> and the rest of the team. Catios, Dean Ball, Sriram, they really all kind of brought it brought the heat.

3:43 >> Yeah, for sure. Of course, the AI action plan was focused on maintaining global AI leadership, particularly over China. And I hate to say it, but we have really been our own worst enemy. You know, it seems like excess regulation on everything from energy production to model development to semiconductor chip distribution. It's really been a bit of an unforced air by the US over the course of the last 24 months and handed a lot of momentum to China. and I think really threatened our leadership. So, this is kind of a 180 to get America back on track. You know, we've underestimated Huawei and the Chinese AI development, model development really at every step of the way. So, I want to kick off today talking about recent developments with the base models and reasoning models coming out of China because, you know, we had this freakout moment earlier in the year with Deepseek that we all remember. Remember, Nvidia stock plummets. Everybody in Washington's talking about it, >> but since then, you know, people kind of forgot about Deep Seek. But the reality is China's been on a roll. I mean, they're dominating the global landscape for open-source models. We've seen six to seven highquality open- source model providers, many of the fastest growing in the world. And this at a time when American open- source, right, Llama 4, has been sputtering a bit, really losing its mojo around the world. Um, so Quen, the open- source model out of Alibaba has passed, I think, 400 million downloads. Of course, that's released like these other models under the Apache 2.0 open source license. So, very open as Bill's talked about. But, Sunny, you tweeted, and one of the reasons I wanted to get you on the pod this week is you tweeted, you know, that the all of these open- source models are really coming together in China. Um, they're leveraging one another. they can distill and generate synthetic data on each other's work. And you went so far as to suggest this might allow them to pass the best proprietary models coming out of the US um yet this year, maybe by Q4 of this year. So why don't we dig in there? What is your theory of the case?

5:56 Why is China, you know, doing so well in open source? And should US model companies like OpenAI and Anthropic be concerned? >> Yeah. So, uh, let let's kind of tie it into, I think, three important things that we see happening. The first one being, you know, the Chinese and and the president addressed this at the AI summit, right? He addressed um the the point around using copyrighted work and he says, you know, he used a great example. If you read a book and you use it, you're not violating the copyright there. And so, he addressed that concern and that was one of the major things that, you know, a lot of people didn't talk about, but I think it's important for the model makers. And so the Chinese just have been able to work around that because of, you know, their position on IP. And what we're really seeing here, and I think, you know, Bill teed it up even better off off of my tweet, which is, um, you know, they're able to compound. So what you're seeing very quickly is both the open- source nature, the open weights nature, uh, allow them to basically compound on each other. So instead of working in silos and instead of having to create giant training clusters um separately they can basically take each other's work build on top of it almost consider it like a remix of someone's model K2 sort of a well-known remix of what what Deepseek had done and now we're starting to see that happen really fast and we're seeing two dimensions of it going quickly. One we're seeing the leading edge models getting quicker and we're then seeing them distill down smaller you know turbo models really really fast as well. There was a a release today of a quen you know 30 billion parameter model which is performing as good as GPT40. So think about that right and GPT40 was you know world class not that long ago. So those are the reasons that we're really seeing an acceleration right now.

7:41 >> You know it I want to dig into this model development in particular. You know a year ago we were talking about these models being these you know stochastic parrots um you know and we really had to compress the entire internet. So you go back to GPT4, you know, and and and you're compressing, you know, the entire internet. But now we really don't need to do it because we've trained them to use tools like the internet, right? The they're true reasoning engines. When I ask a question today, you know, it doesn't just spit out an answer immediately. It goes and uses the tool and it searches the internet. Um, so if you don't have to compress all this wikipedia information, I don't know, take take a subject like World War II, you just need to know how to go out and use the internet to find the information to summarize it in real time. How has that changed the pace of progress and the balance between open and closed?

8:32 >> Yeah. And so it's spot on, right, Brad? what we see now and you see it when you use these reasoning models and what I suggest everyone do is when you're using a reasoning model you can usually expand out its thought process so when you ask a question it'll say oh the person is asking a question about this what should I do let me go and maybe search the internet let me go do a few different things it can have a lot of different tools and so the push has been towards you know really really strong reasoning models and you know we have to give credit there openai really started that with 01 that was really the first reasoning model that was put out there but I think on the back of the research church in the back of what you know everyone talked about and you know the pod's good friend Nome Brown was the leader on that program everyone's been able to look at that and say let's reframe the problem and this allows us to build stronger reasoning models that don't have to compress like you said all the internet's information and once they're coupled with strong tools you start getting these really really incredible results that don't even just show up in benchmarks because none of the benchmarks really allow you to use a tool to answer the results and if they did we're going to see a whole bunch of new set of results that happened there.

9:37 >> Hey Bill, just in many ways I think this validates what you were arguing over the last 6 to 12 months. You said this was likely to happen. Everybody knows you're you're one of the biggest proponents of open source in the world. Um and you were telling people don't make you know the Chinese are going to use open source to their advantage and now we're seeing that in a really profound way. Why is China so successful here and what we what what what can we learn from them?

10:04 Well, I we've talked about it in the past. I won't dwell on it, but China got excited about open source about 20 years ago. It's not a new thing that's happened. And you can imagine when um most of the world accuses you of IP theft that that embracing something like Linux and all the other software open source products um seems very appealing, right? And so I think it became kind of a common way of operating within China.

10:33 Um and it, you know, it's a country that hasn't prioritized uh IP protection the way we have, you know, around patents. And I I could make an argument that, you know, there's way more prosperity if ideas are shared instead of instead of protected. But um I don't know the one thing I don't know is in the in the current AI situation you know was the government promoting open source and encouraging it or did it just develop through competitive forces right >> but but now you have you know a scenario where these companies you know are you know are first of all there's new ones popping up and so you know I almost feel like an idiot when when Kimmy comes out and moonshot and and and And this week, I guess I don't even know how to pronounce it. Zippu, you know, releases a model and I go on Pitchbook and I look it up and they've already raised $1.4 billion. So, shouldn't have been a secret, but I didn't know about it, you know. Um, and all of a sudden they're in the leader tables on Open Router and like, oh my god, you know, they're coming out of everywhere. Um, and what it shows you is just that when you have a competitive dynamic where every single player, and I think there may be seven or eight deep pocketed players with open models in China, they they all learn from each other extremely fast. And and in this case, unlike software, you can use one model to distill the other and make it better. So, it's almost like a accelerated, you know, form of that and you just get massive quick co-evolution.

12:08 And I I came up with a a uh a little analogy for people. I'll try and do it quickly, but imagine you had two communities. They're both farming communities and and let's say there's 10 to 20 farms in each. And in one community, they come into the farmers market once a week and they just compete by selling their products, but then they go back. in the other community when they come into farmers market they also in addition to competing and selling they they they're forced I don't know who would force them but they're forced to share all their best practices from that week >> with everybody and everyone does it and everyone shares their best practices and then if you ran that exercise over two years or whatever obviously the community where the best practices are shared across all farms is going to have a higher global output for the community than the one where you've just got proprietary ideas uh driving you know the individ and and just competition without the idea sharing and that may like even me saying that may cause some people to scream that's socialism or like you know they may not understand open source or how it works or why um but but I do think you end up with a higher fitness level for a community that's behaving that way overall you may end up with a lot less chance of a breakout monopolist like we've had in many of the the sectors in in American technology.

13:32 >> Well, I mean, let's let's just assume the Chinese government is in fact encouraging this in whatever ways, right? I mean, if you look at the release of the AI action plan last week, the Trump administration, you know, had a section >> they did >> which which is about encouraging open-source and open weight models in the US >> saying that these could become standards in some businesses and academic workloads. And it's important they're built on the American AI stack. So, you know, as an aside, the Chinese quickly followed uh, you know, the the American AI plan. I think they released theirs a couple days later where they called for the establishment of a global AI cooperation organization, >> which I thought again is is is is interesting. So, you know, Bill, how do you feel about, you know, like we we we haven't seen that much traction in the US labs uh on open source? Obviously, Llama has been probably the market leader there, but this is for both of you. Uh, you know, handicap for me if you will, how you think this plays out.

14:39 First, Sunny, maybe you start. What do you see at Grock? Do you see a lot of demand for these Chinese open-source models? And if so, what would it take for an American open- source model to catch up? >> Yeah, one of the uh things that we should pull up is the chart of intelligence to price. And one of the things that you see with the leading open source models now, which are the Chinese, is 90% of the quality in terms of intelligence, but at a 90% price discount. And I think whenever you offer that to anybody, you're going to see people want to use that, whether it's individual developers or enterprises.

15:16 And so we're seeing that >> let me interrupt there real quick. So if you're looking at this chart, right, in the top right of that chart, uh you'll see a cluster of these Chinese open- source companies, right? And the vertical axes here being really intelligence, the horizontal axes from left to right being the cost per million tokens. And so you really want to be in the top right of of of that model. High intelligence, low cost. And what this model what what this uh uh chart shows is that you you know to Sunny's point you can get 90% of the intelligence right for 10 or 20% of the cost. And the result I assume Sunny is that you're seeing huge demand at Grock and in Saudi Arabia where you are right now for these Chinese open source models.

16:07 >> We are. And then now just taking that forward. What do people want? They want some accountability. And that's what you'll get. That's what you sort of got out of Linux and say Red Hat, right? As as great as Linux was and Bill was touching on it, the majority of the enterprise was using, you know, a distribution which they could go and, you know, point to someone if they needed something. And so I think the world wants models that they can get from companies that they can go to. And so to answer your question, what happens? I think if we look at Q4 this year or Q1 next year, I'd be willing to say, you know, the a top three worldwide model will be a US-based open source model. And you know, we've got two big efforts happening there. We know we have the open AI open source which you know a lot of people have been working on and open AAI and even you know Sam has commented on that and its release later this summer and then we have all the efforts by meta and if you take you you combine both of those things together I don't think you end up with something that ends up further down the the list in terms of intelligence andor price.

17:10 I um one thing I wanted to highlight about the China situation that I think might inform the US situation. I was um I was having a conversation with this extremely young AI uh entrepreneur that I know and he he he was pouring over the the Zippu I hope I'm pronouncing that right paper and um he asked me some information. I went on Pitchbook and sent it to him who had funded it. And he he he he asked me, he says, "Why is Alibaba funding all these things when they've got their own model?" And because they're in several of the other competitive plays and it reminded me of, you know, a lot of the the the points that I've made about open source is like if you're not if you if you're not confident you're going to win on offense, you want to play defense. And so for any large tech company, commoditizing a a potential threat is is actually quite valuable. You know, you look at what Facebook did with the open compute initiative inside of their data centers. And so, you know, it may just be very well be that that Alibaba just wants to make sure there's no bite dance, you know, that equivalent in the eye space. And that would be pretty rational. And the reason I think that's an interesting data point when you think about the US and you know there there are several big tech companies that seem to be not on the bleeding edge of AI.

18:40 You know you got Microsoft maybe I mean they have access to open AI right now but they might lose that or whatever. >> You've got um Amazon, you got Apple. you know, if I'm at any of those companies, I'd be funding a open-source competitor, you know, rather than funding, you know, like Amazon with Antropic. I think you're in a much better position to encourage open source. And so, I actually think we may >> But Bill, are there are there are there a bunch of open source startup models in the US?

19:10 >> That's where that's where I was going next. I I think you're going to see new entrance um pop up um that that that try to co-evolve with the Chinese models. You know, is Linux American? Is Linux Chinese? Like no one thinks of it as having a a doicile, right? And so um this is just me predicting. I don't you know I I just think you're going to see just like you saw um Kimmy and and Japu pop up. I wouldn't be shocked if you see other new entrance pop up that are trying to be um like sanctioned or or or you know cleaned you know use Redhead as an example Sunny version of these things because I think if you start with access to those Chinese models it wouldn't take you long to move into a near place and you wouldn't have to spend the kind of money the foundational model companies have. Um, and and then you may see a big fight around regulatory capture where they someone tries to say that's not allowed or whatnot. But I I do expect to see that. And if I'm the mistrol team, if you're not distilling on these Chinese models, I don't know what you're doing right now, but I don't have any data on that front.

20:25 >> OpenAI is rumored to be launching their open source model any day. Sunny, what would they have to do? So to your point, you predicted, go back to the intelligence and pricing chart, right? So if open AI was in the top right of that chart, i.e. if they're able to deliver something at the the intelligence of let's call it Quen, and they're also able to deliver it to market at, you know, 20% of the cost. It would seem to me that actors around the world certainly actor it would be uh you know part of the American AI action plan we'd want everybody in the world to use that model and do you believe they have a shot at out competing being the upper right you know by the end of the year and and and if so do you think that will be the outcome like these companies that are using Quen on Grock do you think they would prefer to use OpenAI so long as it was equally capable and equally price performant So two things that we see is brand and you know the US doiciled or you know someone that they can kind of point at that wins and so if that shows up it will win because if you're a company >> and you know at some point you have to you know get your teams to sign off on what is it that you're using what are the risks associated with it and like you know who is liable if something goes wrong and so I sort of feel like um with OpenAI's release and if you know Meta charges back or even if some of these startups emerge that you know we can point at I think we'll see a huge shift back towards those models versus versus the Chinese ones.

22:00 >> Yep. And we really don't know at this point unless you guys have some inside knowledge like when when Meta makes their second push with all these hires that they've made are they are they going to remain committed to open source or even be more open? We don't we really don't know yet at this point. >> Right. You certainly seen some of those rumors out there. Um you know, I've seen rumors that on Twitter that they were debating whether or not they should back away from open source. Um my hunch is that that's a misread. My hunch is that they're going to stay very committed to open source, but they may complement it with a proprietary model. That would be my guess. As opposed to scrapping open-source alto together. And Brad, can I throw one thing out there? I think part of the pitch to get everyone there is that it's open >> because everyone that they're pulling over are coming from closed places. And so if you're really passionate about the work you're doing and you're passionate about where this is going to come and there's only there's only one company that can fund that to that scale and do it open is is those guys. So I think it's part of the pitch.

23:06 >> Yeah. Some of some of the researchers like have a religious belief in it which was evident in the in the interview with the Deep Seek founder like it's it's oddly like higher on their Maslov you know hierarchy than the money. >> Of course they're getting the money. >> Of course it's both. This is I I I think a really important point I want to come back to Sunny is that you're seeing massive demand for these Chinese open-source models today precisely because enterprises around the world can utilize them. They have 90% of the capabilities at 20% of the cost. 10 or 20% of the cost, you know. So it turns out if you deliver something really powerful and really cheap that's more important to these players than American values aligned. But if you gave them something that was super powerful and super cheap and aligned with Western values, right, that you think that would be uh the winning formula for an American OpenAI model to top the uh uh the distribution leaderboards around the world. Is that what I'm I think yeah on a place like open router you know where you can see where this is happening I see we we would see it rise to the top in a few days >> u that's music to David Sax's ears because you know clearly in the strategic plan they're worried about Chinese open- source models dominating uh globally and we're we're if you just watch the pace of releases the quality of the releases out of China the cycle time on the innovation in the open source community out of China. It's faster and better at the moment. So, and the only real big development, you know, we may see some of these new startups, Bill, Bill, we may see a reboot out of meta, but the the one that has, I think, everybody really holding their breath and hoping that we see something really capable and powerful uh is this open source model that's been promised out of open AI. Now, of course, Elon says he's committed to open source as well. Gro 4 is a great model is impressive what they uh what they put out there. Any idea uh Sunny about the open source plans out of out of uh X?

25:29 >> Yeah. So I think like Elon's been you know pretty said it on Twitter that they'll always open source one generation back and so while they were on three you know they should have gotten to two and now they're on four. So I think the the thinking is that they will get there. You know, my my only guess would be is that um you know, right now if they were to open source two or even three, it's so far behind that um you know, what's what's the purpose in doing it? It may not even be utilized and you'll maybe end up having to deal with just, you know, a bunch of internet or Twitter FUD. Just coming back to open AI, I I will tell you it's like one of those things that really you rarely see in the enterprise. It's almost like the demand for like a, you know, like a Tesla Roadster or something or or Model Y before it came out.

26:15 Everybody asked for it. It's like that's the model that everybody wants to use right now. And so, you know, we can't wait till it it comes live and and it it it you know, one it's on Grock, one it's all over the world. I think it's going to be a real big one for everyone. >> While while we have you, Sunonny, um you know, there's really just been this explo explosion in the compute arms race, right? Um there was a big debate you were on the pod a year ago. We had with Bill, you know, had we topped out on compute demand, you know, were we entering an overbuild all Cisco 2000? Um and it's really been remarkable. I think now that's very clear. But just a couple tweets out of Elon and Sam Alman the last couple weeks on this compute demand has have really caught my attention. If you look at this one out of Elon talking about the X.AI AI goal is 50 million in units of H100 equivalent. And Clark Tang on my team tweeted something that broke that down. Um, which showed that that reflected something like, you know, 4 million total GPUs and an energy footprint of like 11 gawatt, right? And then Sam Hoffman comes out and and and talks about their deal down in Abalene for 4.5 gawatt and the fact that they were going to come in well above the $500 billion estimate that they had promised to the government. So these compute clusters now that are being talked about being built over the next 5 years, these are massively bigger than what we were even talking about a year ago. And I think they reflect this move toward inference time re reasoning agentto agent interaction reasoning engines you know Jensen's comment on the pod last year that inference was going to 1 billionx and the consequence you know what we were going to need in terms of compute power to power all that. Um maybe just reflect you're in the middle of all this you're building out your own inference clouds around the world. Um, is this a lot of hyperbole and chest pounding or do you actually see uh the dollars going into the ground in places like Saudi Arabia and around the US?

28:26 >> Yeah, I'm going to I'm going to just quantify it with uh with Google for a second. >> So, you know, in the in in in the um in the Mary Maker bond deck, you know, they have a slide there that shows Google went from 5 trillion tokens a month to 480 trillion tokens a month. And they had just put some press out that they crossed like, you know, 800 trillion. And I saw something today. They crossed a quadrillion. And I had to look that up. That's a,000 trillion. And so in a in a course of a year and a few months, they've gone from 5 trillion to a,000 trillion. So that's 200x right there. I mean, that just shows it to you. Um, without having to look at anything else.

29:06 >> Uh, the amount >> every single search query on the planet today is now an inference transaction. Correct. And so um you know and you see it uh you know in anthropic uh with you know their continued fundraisers going you know through the roof uh that's happening because they're they're seeing the amount of token consumption. And so anywhere we lay infrastructure um we fire up a rack it becomes fully consumed within a few hours. >> You're talking Grock.

29:37 >> So you have demand far outstripping supply even at Grock. Yep, we do. >> So, Bill, you see these fundraising announcements that are being discussed. Just CNBC's reporting tonight. I think that Iconic is going to lead a $5 billion round into Anthropic at 170 billion. In the case of Anthropic, it's, you know, uh 170 billion on rumored $5 billion in revenue. X has been rumored to be raising at 150 to$200 billion. So, you know, you've never in the history of venture, you've never seen fundraisers like this. One of the topics that is being hotly debated on on the Twitter, uh, is that there's massive intervening delilution in these rounds, Bill, because of the employee option grants or the employee RSUs that are needing to be granted to keep the employees uh, you know, uh, in these businesses. So just observing this a little bit from afar. I don't think you guys you're a direct investor in in any of these uh major labs.

30:43 What what do you see? What do you observe? What are your warning signs uh you know about the size and and the demand that you see in these rounds? >> Well, I yeah, I've never seen anything like it. I mean, I saw you know, through the Uber andyft situation, I saw a precursor to this, but this these these dollars are even bigger. and the amount of money that these companies are willing to lose in a year. You know, it's I still think it's particularly interesting that Google has to compete with Open AI because OpenAI is going to lose 7 billion this year and Google won't like like they would never allow themselves to do that. And and this started back in that previous era where where for the first time ever you saw private companies have a competitive advantage and that they can be more risk-seeking with capital than the public companies are allowed to be.

31:34 >> Um >> but I think in the past 12 months we I we've seen um OpenAI Meta and certainly X you know move into this place. I call them I call them the cost is no object the CNO group um where they're just they're just putting out press release after press release and opening data center after data center and you know we there there are other people I think you know you look at Microsoft you know choosing not to extend their capex budget you look at that Amazon example when we spoke to the the Levant brothers at codeu where they're they're not keeping up with the the Nvidia purchases relative to their AWS share. And so there are a few companies that are back on their feet and there's a few companies that are really pushing the gas pedal. And I can't there's a few in the middle. I can't tell um whether Anthropic has the uh audacity and the means to raise enough money to to start building data centers themselves. They haven't so far. Um but it's uh you know it's a sport of kings, you know. It's it is a sport of kings like there is um never been this amount of money spent right now. Nvidia and others building in the stack like like uh Dell that we we spoke to a few weeks ago like they're the winners um in a pick and shovel game that's got this amount of aggressiveness. It's uh and maybe maybe SK Heinik too. I don't know.

33:06 Sunny, Sunny, do you see um you know going back to the well-worn cliche that every glut or that every shortage ultimately ends in a glut? Um do you have any evidence on the horizon where you see uh uh supply outstripping demand? >> No. I mean I I was going to ask this to Bill as he was just saying it. Bill, like on a daily basis, are you consuming more tokens or are you consuming more, you know, just traditional um you know, web lookups, right? And I'd be willing to guess you're consuming more tokens.

33:42 >> Oh, it's insane. >> And and and tokens are are increasing like when you're using those reasoning models, you don't see all the tokens, right? They don't publish it, but there's Right. This is where totally Yeah. 10 to 100x more than the your very first AI search. Yeah, for sure. Absolutely. The one thing the one thing I will So you're right. I I'm doing >> Go ahead. You finish. >> No, no, no, no, no. And I just wanted to say even yesterday, Anthropic had to put this uh you know, press release plus, you know, product change out saying, "Hey, we've got to throttle everybody, right?" And this is like cuz we have all these people using way too much of our our services. So I think you know if you have intelligent models right and you have the capacity for it it's one of those things people are consuming Jven's paradox or whatever. Yeah. Sorry. Go ahead, Bill. You're >> No, I was just going to offer one caveat to to this uh super uh exciting line of of of thought, which is because of the amount of venture capital out there. Um companies are not pricing the cost like no one no one none of the model companies I don't think anyone even in the verticals like no one's pricing the cost because they're pricing to take market share. And you know, you and I, Sunny, we had a previous discussion about the unlimited pricing and inference has variable costs. So, is that even sustainable, right? And and even when people say to me, "Oh, well, we're going to run out of power." You wouldn't run out of power if you just took the price up. Like, like the thing that throttles demand is price, but no one here is raising price. Like Anthropic doesn't need to throttle. They just raise price, but they're not willing to do that because they're afraid to lose share. And so everyone's pricing to share. That means they're pricing under cost. There's rumors of even some of the best known brands in AI having negative gross margin. And so I don't know when that settles out, but that there that will create a bump in the supply demand curve if that ever has to be, you know, fixed. Um but for now it doesn't >> because because >> Can I ask a question there?

35:55 >> Yeah. >> Yeah. No, no. which is going back to the point that we made on open source, but if you know in the back of your mind that there's something that's 90% as good, but it's 90% cheaper. How does that factor in because we've also never had that that factor as we're going through this growth curve. I mean since almost since we started the pod you know I've I've routinely highlighted that the steepness of that price curve on you know as you as it kind of becomes you know less less cutting edge is something I've never seen before. I've never ever seen it. And I'm sure that a lot of um people sit around and say, well, it's okay if I'm losing money here because, you know, six months from now, I'll just use the older model. And we also talked in the past about how in the internet age all the startups began with Oracle and Sun and eventually they all moved to Linux and MySQL. And so there was a there was a we got to win at all cost phase and then there was a phase where you started worrying about cost and optimization. And so one day, one day we'll likely, you know, make that make that move. And and and I a few of the companies I've talked to that are running inference at scale, they are already starting to think that way. Like they're looking right they're looking at it, you know, from that from that lens.

37:20 But I but >> well and and I think that's why Grock and Cerebras are doing so well. But Sunny, give us give give us an, you know, an an example. I would imagine that the wind surfs of the world and the cursors of the world and all these folks who are building these you know these coding agents they've got massive you know demand for their applications right but they're paying through the nose to anthropic or to these underlying proprietary model providers to be able to do that so what's the dynamic that you see there do you see them running to implement Quinn or some of these cheaper models >> yeah without kind of getting into specific any one of them, but like you know multiple folks are building their own models uh based off open source and >> right so they could just go distill any one of these models.

38:10 >> Correct. >> Right. And you know given that they're these very uh lenient Apache licenses and eliminate the entire cost of sonnet that sits under it for 70% of use cases. >> Yeah. And and like Bill said, turn that into a premium offering, right? and say that's the you know the gold and you know the silver and the bronze are built off uh you know something that's like I said oneten the price >> that seems to me Bill you know if if I had to forecast you know if I'm open AI I'm running a consumer business with really high gross margins right because consumers are uh uh less sensitive to what they're paying their their their their intensity of use is lower. Whereas when if if somebody's writing code, you know, there's the variable intensity is high. And so, you know, for them, it would make sense to launch an open source model back to where we started the pod and price it really low to, you know, drive share knowing, you know, it reminds me a little bit of Amazon back in the day.

39:20 Amazon had this monopoly retail business they could use to subsidize AWS gain share for a decade and then begin to take price. That would be a rational strategy for open AI to follow. So you take the profitable consumer business you use it to subsidize you know the the the market share in other applications that you hope to build. >> Certainly a reasonable strategy. I I I I I'm not inside that company. you have way more knowledge than I do, but um you know, I pay the $200 a month and I do every one of my searches on four or five and I'm probably I'm probably negative, I would think. Yeah.

40:00 >> And so I do think there'll be some rationalization where these models um kind of self-pick which one they're running based on what you need. I probably don't need to be using >> and move to more of a consumption logic. Yeah, >> move more to more of a consumption logic. they're already uh they're already doing that in in parts of their enterprise business. Um and and you know, as they've transitioned to more of this consumption logic, I think it's led to some real unlocks. Uh you you know, for the business, I think it that makes it harder if you're in the the lab game and you don't have a consumer product and you don't own an application that you can drive high gross margins. I think then gets back to this question, how long can you run the your business for share, right? Hoping that someday because you know they all exist at the beneficence of the capital markets and the capital markets are willing to provide an incredible amount of capital to these businesses today. They sure are.

41:00 >> Um but but but you and I have we've lived through these periods where that disappears quickly. And can I can I put something there just to kind of hear your feedback on it guys which is look at Google and the TPU and Google is clearly you know by these numbers that we're seeing right putting out more tokens than anyone else right do they do you guys believe they have a strategic advantage because they have their own hardware they're not having to pay you know 80% margin on something that they can generate tokens with like how do you guys look at that business and say clearly they look to be sort of the largest at least openly saying um the largest processor of tokens. I think look the key data point in that case which I don't have the data but I'd be glad to repeat it if someone shared it with us is how many non-G Google uh applications are running on the TPUs like how many third party customers are using because I what I've heard or what the you know the general perception is is that that most of their proprietary TPU transactions are their own applications.

42:05 >> Yep. But that's probably where most of their tokens are being processed anyways at this point, Bill, right? Like transcribing YouTube videos and you know all in Google Meet you can turn it on and all the searches. >> I I was just inferring in your question, maybe I shouldn't have been that that they'll have an advantage for Google Cloud. And in in order for that to be true, they need to to have this crossover moment. One quick thing, Brad, on on Open AI, you know, I'm I've I've been writing a book, which I've talked about frequently, and I've been quite uh although I guess there's some privacy things now you need to be worried about, but I've been quite open with Open AI about the book and doing research, you know, along the way. it knows a tremendous amount about my book right now and I can ask follow-up questions without having to put the whole book back in the prompt again um because of that. And so I can I would I continue to believe that OpenAI's um most likely chance to long-term um success comes from switching cost and lock in more than it will come from staying on the edge of the of the model race because I think >> and the pricing and the pricing power that comes with that brand dominance, right? No doubt.

43:24 >> Because the the fact of the matter is you said you're paying 200 bucks, you're getting more than $200 in value. I don't know what the price is, but I know that if it was variable, you would pay a hell of a lot more money to use that service. >> I would I would but but but the the the lock in once of these systems starts to truly understand you and have all your historic knowledge um I think the switching cost will be very high at that moment in time.

43:51 >> Sunonny, back to your question about the TPU and the advantage of that vertical integration um you know to to Google. I think it's it's it's too early to know. What I would tell you is that uh they they've absolutely made some changes I think over the course of the last you know 3 to four months to accelerate the business. You you've seen the news about OpenAI uh leveraging TPUs for some of the inference uh you know uh demand that that they have. Um ultimately what I've said all along about Google is there many ways the best position company in the world but a lot of their advantage right is no longer much of an advantage right and namely that you know they were the dominant place where the consumer started every single query and we know today that's just not true in you know when people are looking for answers Bill's book is not in Google bill's book is in you know chat GBT and that's the >> actually it's in Google doc but but you're right you're Right. The knowledge of it, the knowledge of it, >> the knowledge of it and the interaction and the the token generation. So my only point is this, >> the battle for the consumer is ultimately where the value occurs, Sunny, not who runs, you know, what hardware. And so Google's dominance, the it's been the greatest b business in the history of capitalism for 20 years because they owned the consumer. They owned the verb in something that was extraordinarily high margin. And all I would say is the first real threat in 20 years came about because in the chat GBT moment it's continued to accelerate. I think chat GBT will cross a billion weeklys like maybe this year. Um you know probably this year I would guess. Um, and so that to me has always been the case, right, for open AI. And when you look at the rest of these frontier labs, the case you've made throughout this pod, about seven of these models in in in in China, being able to open source, distill off one another, drive up intelligence, drive down cost. What that tells me is the model layer is being increasingly commoditized and that there's not going to be a lot of intrinsic value in in in that intelligence layer, that operating layer. You're going to have to build applications that guys like Bill Gurley and and you and I are using every day.

46:09 And that's where the battle is on the consumer side. You're going to have the exact same battle when it comes to coding agents and general enterprise applications. And I've said there, I think it's going to be more of a uh uh you know, heterogeneous world. I think they're going to be lots of players that compete. I think the margins will be lower. Yeah. >> Uh in that world, but it may very well be that the tide is going up so much here, right? The whole world is transforming so much around this that you're still going to have lots of players who do incredibly well. I have to say I'm surprised if you would have told any of us a year ago or 18 months ago, right, that the combined enterprise value of Open AI and Anthropic together would be over half a trillion dollars.

46:53 And you throw X.AI in there, it' be a trillion dollars, you know, across the the three of them roughly. um it's bigger and faster and the compute demand is higher than any of us I think anticipated. Sunny, hopefully we can keep you on here for a bit. We're just going to wrap up with a um you know a topic that I think has dominated really the conversation in the markets over the course of last year and that's been about tariffs and kind of the reordering of global trade. Bill, you I know you had some some questions, some thoughts about it and I'm I'm happy to dig in and talk about it as well. Well, I mean, I would I would really just love to hear from you, Brad. the the the markets got very nervous when the um what was it liberation day when when the kind of unpredictabilility of how big some of the numbers were and what that might mean and whether we were walking away from the notion of comparative advantage and and I think the markets got spooked and you know I think you turn around and look at where the markets are today and and we've really gone through an evolution of how the Wall Street is interpreting the both the initial um launch of the tariffs and the reality of where they're landing.

48:12 So, how would you how would you describe that? And and why do you think the markets are getting um very comfortable with the where where they're landing? I mean, not only comfortable, we're at all-time highs, >> you know, and and April 2nd, I was uh going, you know, on CNBC saying the nuclear Navaro was going to be a disaster and I'm out, right? And that's this the the the amazing thing about this administration is there's really uh you know it's a team of rivals within the White House and you had Besset and Lutnik who were basically outlining this plan for you know let's call it 10 to 15 to 20% tariffs across the board that would amount to about 300 billion in total tariffs up from 75 billion uh in 2024. But you had Navaro who was basically saying we're going to replace the Internal Revenue Service. We're going to get rid of the income tax and we're going to have two trillion of tariffs. Okay? And I was very clear and I think the market was very clear. We all voted with our wallets and we said until the president tells us whether it's door one or door two, we're out.

49:17 >> The market shot first and asked questions later. And that's where you saw that huge draw down in the market. The NASDAQ was down 21% right at its trough this year. Now the NASDAQ's up over 10. It's a 30% move in about 60 or 70 days, which is extraordinary even by the historical patterns that we've seen over the course of the last 5 years. But let me back up here for a second. >> I think the consensus view of all economists, right, 90% of economists is that tariffs are going to be bad.

49:52 They're going to be a tax that gets paid by the US consumer. There was a small group led by you know Scott Bessant and Kevin Hassid at the National Economic Council that said no it's actually going to be different this time and the reason it's going to be different this time is their theory uh argued was that the world had become dependent upon exporting to the United States so that the total trade deficit to the United States of goods and services was about 915 billion last last year a $1.2 2 trillion goods deficit and that basically meant that China was selling a lot more to the United States than than they were buying of US goods and and so what Bessant and Hasset postulated was that these countries have no choice if we impose a tariff on them so long as it's not draconian 70% 80% what Navaro was talking about if we impose a 15% tariff on them they have to eat it the producers have to eat it because otherwise they're going end up laying off millions of people in Vietnam, in China, in these countries.

51:01 And politically, they can't afford to lay these folks off. So that was their theory of the case. The consensus economist said, "No way is that true. You're going to see massive inflation." But what have you seen? You have not seen the in the the inflation percolate through. I will caveat this by saying yet. >> Okay. So here we are in July. The e the consensus economist said it would have already happened and the national economic council was out with a paper last week that deconstructed core PCE.

51:34 So that's the uh uh uh what the the best proxy the Fed watches for inflation since the start of the year. And it showed this was really interesting. Import prices have been going up at a slower rate than domestically produced goods. Okay, so this is the exact opposite of what you would have expected from tariffs. Of course, you would have expected the imported prices would have been going up more than domestic uh goods. And so we'll show these charts um and we'll we'll put the link to this paper. People ought to take a look at that. But to me, um, when I look at the president's, uh, the deals he's landing, the deal he just got announced yesterday with the European Union, 15% on all goods coming from Europe, 0% on US goods going to Europe. So, an opening up of Europe markets, paying us 15% and on top of that getting commitments like $750 billion, almost a trillion dollars of energy purchases from the US. Or look at Japan, which they announced last week.

52:39 Another huge market. Again, similar. They're going to pay tariffs to the United States. No tariffs imposed on the United States. And they're going to invest $550 billion into the US in a way the president gets to direct. So, I would I would I just think we need to give the president credit where credit is due. Everybody said this was going to lead to retaliation, to trade wars, was going to be disastrous for the US. And all we've seen so far is deals, deals, deals, deals. And I have to say, if this was the CEO of one of our companies, right? Let's say we had a board meeting at the start of the year and he outlined these plans and we said, "Hey, we're really nervous about this. You know, this is a high-risk, highreward strategy. it's either going to backfire and we're going to fire you or it's going to work really well and we give you a bonus. If we're measuring them halfway through the year, I would say that he's in line for a bonus based upon the trillions of dollars that are going to be coming into the United States and the fact that we now have a 300 to a $350 billion recurring, you know, stream of revenues into Treasury in the form of these tariffs which are being paid. and and and I think Bessant said last week in the month of June we had our first surplus in the United States monthly surplus since 2015 >> right because of the tariff revenues that are coming in. So I will say this I was on the fence I knew the nuclear tariffs the trillion or two trillion I knew that was a disaster. I said if we landed the plane where Bessant wanted to come in at 300 billion, I thought there was a decent chance that those prices could be passed on in the home countries and so far it looks like that's the case.

54:27 >> Do you have any concerns? What's the uh anything you're watching out for? Well, I think the the number one thing is core PCE h did bottom last year and it started to tick up and so we have to keep our eye on inflation. And of course, I think it's almost impossible to conceive that we would have totally reordered the entire global trading system on the first pass with zero mistakes. So, we're going to have some some goods and some products that, you know, consumers are US consumers are going to end up paying the taxes and we're going to have to go back and fix some of these things. But, I will say that it's turning out massively better than consensus criticisms. I mean, remember Larry Summers at the CO2 event.

55:11 >> I mean, this was just a few months ago and he was saying this is the biggest economic disaster of of his career. And I I don't think you can describe it that way. The markets are the voting machine is telling you and these are a lot of sophisticated investors. The voting machine is telling you that no retaliations, no trade wars, all these deals getting done um uh works for the US economy. And I will tell you that the Atlanta now Fed tracker which tracks real-time GDP has now ticked up back to 3%. So after going down a lot in April, it's bounced way back up. So economic activity uh appears to be going up. And then one final thing here, remember one of the key reasons for doing this, right? Wasn't just because we were, you know, there was I mean the EU conceded in the trade negotiations that they that our relationship was unfairly balanced in the direction of the EU. I mean they said that's the starting point. So we have to rebalance it. But on top of that, a key reason for doing this was to support the domestic production of critical national industries and to make our supply chains more resilient chips, data centers, energy production. At Altimter, we just led the series A in a company, I don't even know if we've announced it, but I'll announce it here, which is an all-American producer of rare earth magnets. Right? So these are now viable investments because of the tariffs and the resolve of the government to reonshore these critical supply chains.

56:38 So I mean that's a huge benefit that you would be willing to pay something for but we're getting the benefit and on top of that we're getting paid. >> Yeah. Can I add one thing guys like not on the economic side but like you know running the supply chain you know Grock and look we're we're fortunate that like the majority of our supply chain is US- ccentric including our chips but >> um you know we still have small discreet components and Brad exactly what you were saying is happening is that the the producers the manufacturers of those things are coming and you're negotiating and we've had pretty you know significant negotiations with those folks and that that was I think like not like you said wasn't anticipated. The other thing is it hasn't been um static.

57:18 The one thing again I'll go back to this administration they moved like we've we've seen you know our tariff schedule move around probably six to eight times since this all has started because they keep evaluating they understand they listen and I think that's also a function of how this administration operates and we want to give them credit for that like people can go and share with them hey like these things are not available we can't replace them right away and I think that that's also helping with that last thing you said with the investment you're making is companies get established on shore to take advantage of what these tariffs are causing.

57:51 >> Sunny, has your supply chain planning and that that dynamic nature, is that started to settle down? Do you see this, you know, are we reaching the end of the tariff negotiation such that everything can kind of settle down and operate? >> It's gone from week to week or even day to day when it first started and trying to figure out what happened to like now we're looking at it monthly. So, it's definitely settling. And Brad, we still have a big thing looming with a China discussion, correct?

58:19 >> Yeah. You know, listen, so we landed Europe, we've landed Japan, we're going to have, you know, the long uh the long list is is is going to be coming out, but when you look at our big trading part partners, the EU, we do, I think, $900 billion of trade with them a year. With China, it's about 600 billion, but we have about a $300 billion goods trade deficit with both Europe and China. So they were the the the two big ones. Um China is the big enchilada because it's not just trade with China, right? It's strategic. It's national security and its trade. Um and it's the AI race. We know that, you know, the rare earth ban on Chinese magnets was devastating to US industry. We know the retaliation that, you know, where where H20s were cut off in terms of Nvidia's chips back to China. reading the tea leaves. I'm going to I'll go out on a limb and I'll say um the consensus still believes like that the China thing is going to be a problem or that it will be small. I think this president wants to do the biggest deal ever done with China. I don't think he's dogmatic at all. I don't think he's some big China hawk. He said at the AI summit last week, I'm a deal junkie. I like to do deals. You can't be the biggest deal maker in the world without doing a big deal with China.

59:41 >> All right. >> Right. And I and I, you know, if you just look at today, the Chinese uh reciprocated in a way I think the the US government was looking for. They said, "Hey, we'll postpone all of our retaliatory tariffs." They've invited the president to China. The president has suggested he's going to go to China the first week of September or sometime between September and November. I think there is a very very big deal that's going to get done with China that's going to reorient the relationship in a big way. And let me just like tease this. The president said earlier this year something that caught all of our attention. He said, "You know, if I could wave a magic wand, I would cut the defense budget in half for the United States, for China, and for Russia."

60:26 We've never heard a US president in history utter anything like that. >> Be amazing. That is a what I would call an extraordinarily flexible mindset. And if you go into this deal negotiation with China with that sort of flexible mindset, I think it could include all of the above. Rare earth chips, maybe even military cooperation, certainly a rebalancing of trade. I think China's, Listen, we entered the year with China paying 15% in tariffs. That was pre-Trump. They were paying 15% to the United States in tariffs. So I don't think we're going below 15%. I think they're going to pay >> that in in Trump for >> Trump won and Trump won, right? So I think that they will continue to pay at least 15%, but I think it's going to be much more structured, much more nuanced.

61:12 You know, Besson's been very clear China has to rebalance to domestic consumption and away from an export economy that's really uh sticking it to the US in terms of the trades def trade deficit. I think China gets that. I think they want that for their own country. I think they're willing to do that. I think the United States understands that can't happen overnight. It has to happen over a period of years. I think that there's going to be a big Chinese deal done before the year's out.

61:40 >> So So let let's close with this, Brad. You've you've often on the pod um been willing to speak about your your own temperature for the US markets and and and where you whether you're net long or net short. and you >> you've on this podcast been more enthusiastic than I've ever seen you both about uh about AI, but but this this China theory that you have, I think, would would cause the markets to rip, if you're correct. Um, but you also said we're at all-time highs, you know, and so you want to buy low and sell high. So, where where's your head hanging?

62:18 >> Uh, we did a pod, I think, around May 2nd. Well, we did the pod in March and I I I said we're out of the market. I remember. >> Right. And you and you said you're early and liberation day came and we were happy to be out of the market. On May 2nd, I said we're all we're all back in because the Bessant consensus has won. It's going to be 300 billion. We're going to land the plane. And I outlined a flight path. I said you can land the plane, no inflation, you get rate cuts and you know, and it's kind of off to the races. And we're up 30%, you know, off of that bottom in the NASDAQ since then. So 30%'s a huge move, but when when you when you telescope out, we're up 10% for the year.

62:56 >> Okay? And if I had told you guys on day one of this year, here's what's going to happen. >> We're going to rebalance global trade and we're going to land the plane around 300 billion. We're going to have the economy grow at 3% accelerating. We're going to um uh you know have no inflation heading toward rate cuts by the end of the year. um you know that is the backdrop you know that I and we're going to have all of this AI demand and accelerating demand for AI compute I I would have said the market can be up at least 15% for the year I think we've captured a lot of the return for the for the year Bill I will tell you this though we see tons of opportunities and so I would say that we're also bullish on what we see happening in AI Sunny's going to raise a huge new round here we're happy to be investors with Sunny as well and and it's extraordinary to Watch watch what Sunny's uh No, seriously. He >> wasn't supposed to disclose that. He can edit it out.

63:53 >> He has He has it, but his face is turning all red. >> Have a good week. And >> thank you, Sunny, for coming on with us today. Thank you so much. I know it's >> Thanks for having me. Thank >> Yeah. Thanks, guys. >> Byebye. >> Take care, guys. [Applause] [Music] As a reminder to everybody, just our opinions, not investment advice.