Anthony Pompliano
The AI Boom Is Bigger Than Anyone Thinks
🇬🇧 EN🇪🇸 ES
Ask the tutor about this video
Sign in with a verified email and ask the tutor anything about this video.
TL;DR
Research summary
▶ Tech competition and interview setup
Ives opens by contrasting his experience seeing fabs built in Taiwan and Korea with an anecdote about a fight at a New York airport, linking the contrast to his concern about US education and competitiveness. He says the US is ahead of China in tech for the first time and argues that this should continue. The host introduces the discussion of AI investments, private model labs, regulation, personal assistants and Ives’s new IVAI fund.
▶ AI spending and the broadening trade
Asked where public-market value lies, Ives says monetization by Microsoft and other hyperscalers in the prior quarter marked an inflection point. He sees AI investment spreading beyond memory into infrastructure companies such as Dell and Cisco and cybersecurity. He estimates that less than 15% of the next 3–4 years’ spending trend has occurred, describes total spending as $4–5 trillion, and says each dollar of capital expenditure could have a $5–$6 multiplier across tech. He expects near-term ebbs and flows from regulation and bearish commentary, but believes a multi-year tech bull market is beginning. On regulation, he says technology tends to move ahead of regulation and argues that some calls for restraint may amount to regulatory capture, while acknowledging the industry’s public-relations problems. In his view, slowing US AI development would let China accelerate. He says leading labs such as Anthropic and OpenAI have an advantage in models and enterprise, though the gap with Meta and Google is narrowing; he expects models to become more commoditized and sees data as a key source of value.
▶ Model economics and sovereign AI
The host argues that large model labs may face more pressure than investors expect: token consumption is supposed to lead revenue, but he believes revenue has plateaued, particularly at Anthropic, as users shift toward specific applications and their own intelligence. He describes his company’s progression from ChatGPT and Claude integrations to reducing model calls and building custom models and infrastructure. He says this has improved performance and accuracy while lowering cost and latency, and that the company still uses external models but less often. Ives agrees that sovereign AI and control of company data matter, while arguing adoption remains very early: he cites 4% of US companies as having taken the AI path and says advanced adopters are beginning to pursue sovereign AI. He sees substantial runway in enterprise adoption, consumer products, and markets including Europe, India and the Middle East, and expects another two to three years of strong growth.
▶ Regulation, open models and IPO prospects
Ives says the White House meeting was intended to align tech leaders amid political debate and concern that negative, dystopian messaging could lead to data centers being shut down. He favors self-regulation and says shutdowns would benefit China. On open versus closed models, he credits Nvidia’s Jensen Huang with understanding the need for open source to reduce costs and broaden access; Ives also praises Alex Karp’s global and enterprise perspective. The host highlights a potential squeeze on model revenue from falling token prices and consumption per customer. Ives expects an Anthropic IPO, possibly after the midterms, and argues that public listings of Anthropic, OpenAI and SpaceX could benefit the market. The host cautions that strong products do not necessarily make strong businesses and cites potential competition and fragmentation, including a new model gaining use where frontier models were too expensive. Ives counters that Anthropic and OpenAI have built enterprise sales forces and says low current AI adoption leaves room for growth. He also points to Meta and Apple’s large user bases and argues the AI trade is spreading to traditional tech, infrastructure and energy.
▶ Portfolio allocation across AI beneficiaries
Ives says investors should look beyond the direct AI leaders to second-, third- and fourth-order effects and keep a multi-year horizon. He recommends owning chips and hyperscalers, citing an Asia check that showed 13-to-1 chip demand versus supply, and names Nvidia, AMD, Microsoft, Amazon and Google. He rejects the “SaaS apocalypse” narrative and points to Salesforce, ServiceNow and Palantir, while saying Palantir could become a category changer as free cash flow rises. He also sees opportunity in cybersecurity companies including CrowdStrike, Palo Alto and Zscaler, as well as infrastructure and energy. He advises holding core names and themes while anticipating transitions such as Tesla moving from electric vehicles toward robotaxis and autonomy; he says investors may miss the move if they wait until that transition is complete.
▶ Consumer AI and the macro backdrop
On personal AI agents and consumer products, Ives sees a new investment area and says infrastructure providers may benefit across enterprise and consumer markets. He expects a consumer advertising opportunity to develop and says new categories and investable companies may emerge over the next 18–24 months. He believes chip supply and demand will not reach equilibrium until late 2028 or early 2029, and expects tech to lead despite oil, geopolitical risks and other macro concerns. The host lists potential headwinds, including interest rates, oil near $100, the 10-year yield above 5% and falling home prices. Ives argues that the market has held up better than a bearish scenario might imply, attributing elevated yields in part to growth, debt issuance, data-center construction and capital expenditure. He says bearish narratives can create investment opportunities.
▶ Sponsored segments
The interview pauses for three promotional segments. Arch Public advertises automated strategies across stocks, commodities and ETFs, plus a crypto tax-loss harvesting tool. Lava advertises a card offering up to 5% back in Bitcoin and other financial services. Investor Health advertises telehealth programs and compounded medications, with the ad stating that compounded medications are not FDA approved and that treatment requires evaluation by a licensed provider.
▶ IVAI and access to private AI companies
The host describes Ives’s new AIV Ultra Fund, ticker IVAI, as a publicly traded permanent-capital vehicle with exposure to private companies, and says it raised $200 million. Ives calls it the first public company to invest in private AI tech companies. He says the aim is to give public investors access to leading private tech companies and identify potential future leaders, rather than limiting access to a small group of private-market investors. The host agrees that it reflects the blending of public and private markets. They discuss the value of data businesses, including the host’s investment in MicroOne, which pays experts such as doctors and lawyers to create data, and an inquiry about licensing podcast data. Ives calls data a major asset and says companies’ desire to control data is central to sovereign AI. He describes company adoption as an evolutionary process of trial and error that may create opportunities for more businesses.
▶ Sovereign AI, productivity and closing views
The host says routing queries to in-house models on company hardware has produced about 97% cost savings compared with frontier models in his company’s experience, and claims its tailored system is more accurate on tax, mortgage and credit-card topics. He attributes performance to application-specific use, datasets, evaluation methods and reinforcement learning, and argues that large companies may build their own systems. Ives says data sovereignty makes the market broader than a duopoly of OpenAI and Anthropic, and describes the AI buildout as being in its third inning. He cites a projection of 800,000 data centers built in the next 12–18 months, even if 10–15% are rejected. The host describes learning to build software with AI and says this blurs the boundary between technical and nontechnical workers. Ives argues AI will create more jobs than it removes over the next decade; the host says agents have replaced some hiring at his company while supporting faster growth and later hiring. They conclude that investors should avoid getting swept up in narratives and should consider adoption and productivity gains. Ives directs listeners to his online profiles; the host reiterates that IVAI is the fund’s ticker.
Full transcript
[0:00] For someone like myself, I've covered tech my whole career. 3 and 1/2 million air miles, so many times like in Taiwan, in Korea, you see them building fabs 18 hours a day with one bathroom break. Then I land in New York Airport, there's a fist fight at the Dunkin Donuts, and you wonder why we're 17th in math. Now, for the first time, US is ahead of China when it comes to tech. You don't want that to end. [music] >> What's going on, guys? Today, we've got Dan Ives on the podcast. Dan is the partner and senior managing director at
[0:31] Yorkville Ives, and he's here to talk all about the AI trade. We talk about the sectors that he's interested in, the sectors where he thinks value's going to accrue, and maybe some areas where you should be cautious. We talk about the private AI model labs, where is pressure coming from, or is AI actually going to kill us all, what happened at the White House meeting, and then we even get into what's going on with all of the personal assistants that have been coming to the market. Dan's got some very unique thoughts on the AI trade, and I think that if you watch this, it's going to make you a smarter investor in the market. And then Dan finally finishes up, and he talks about the Ives Ultra Fund. It's a brand new fund that he's
[1:01] brought to market. It trades under the ticker IVAI, and I think people are going to be very interested to see how they can hold a public stock and get exposure to the leading private AI companies. Here's my latest conversation with Dan Ives. All right, Dan. AI trade, obviously, was super hot earlier this year, then got a little cooling off in the summer, but it feels like it's starting to come back here a little bit. And we have the private companies that are knocking on the door of going public in Anthropic, Open AI. But the public stocks, where do you see value through the end of the year? How how are you thinking about the AI trade today?
[1:31] >> Yeah. Look, I think the demand, I think it started last quarter. That's where it changed. Cuz the monetization you saw specifically with Microsoft and the and the other hyperscalers, I think that was an inflection point quarter. >> [snorts] >> I think now a lot of it is on the second, third, fourth derivatives. And then, of course, memory stocks were there. But I think you're seeing it across infrastructure with Dell, with Cisco. I think with cybersecurity stocks, you're starting to see it play out. Look, take a step back. We're less than 15% through what the spending
[2:03] trend's going to be the next 3 to 4 years. It's 4 to 5 trillion dollars. For every dollar of CapEx, there's a $5 to $6 multiplier across the rest of tech. So, I think we're going to go through these sort of ebbs and flows, the regulatory capture, the Dalio essay. I mean, some of these things will continue to sort of get in the way to some extent in the near term of the trade, but I just continue to believe we are in the early days of what's going to be a multi-year tech bull market. The first time in 30 years, US has added China when it comes to tech.
[2:33] >> When you see Dalio come out and basically say, "Hey, we think that there's a 5% chance or 10% chance of AI killing us all. We must slow down." And, you know, kind of all of these calls for essentially like, "Hey, government, regulate me harder." I've never heard a business that is serious talk this way. What's your general read? Are they sincere in this? Are they just trying to use it to slow down their competitors so they can catch up? Or or what's going on? >> I mean, 1910, the guy who said, "We can't go down this car route. We have to stay horse and buggy. These cars kill
[3:03] you." I'm I'm just saying, the it's technology. Technology always goes ahead of regulatory. And, I think some of it is I'll call regulatory capture. Like, you go to the top floor and then you pull the ladder up. Mhm. So, I think there is some of that going on. And, I think that's also why you had to have that sort of all hands on meeting, you know, with Trump at at the White House because you have to get everyone on the same page. Cuz the reality is going into a midterm election cycle, you know, >> [clears throat] >> talk like that, and me and you have talked about this before. I think a lot
[3:33] of the tech industry, they created the PR nightmare, the black eye that now they're contending with. But, it's it's an arms race. And, if we slow down, China accelerates. That's not a debate. So, that's why every time I in terms of like the politicians in the Beltway, many, you know, do you want people that are still using Blackberries and flip phones to ultimately be the ones that are, you know, the direction of innovation in this country? And I think that Look, it's not going to slow down. And I think you saw how stocks reacted.
[4:06] I think if you go back I go over the weekend the bears that have called 10 of the last two downturns. They're like, "Oh, my cuz remember the bulls are usually watching football hanging out with their family." And the bears on Saturday, that's usually where it peaks out. But then the reality is is that you you saw what Jensen talked about, what Nadella talked about, what Karp talks about in terms of sovereign AI. That's more what I view as like a very sort of middle-of-the-road view, and that's my opinion. And this is not slowing down anytime soon.
[4:36] >> What I found interesting was, you know, right after the call for the slowdown or pacing or whatever the hell they were saying, um Zuck came out and Zuck did not talk about slowing down. Zuck talked about, "Hey, it's our job to release products that are safe." And my general framework, but I want to hear if you agree with this or not, is the companies that are losing money, that have negative cash flow, they're the ones calling for the slowdown. The companies that are profitable like the Facebooks, the Googles, etc., I don't hear them calling for a slowdown at all. And it almost feels like depending on whether
[5:06] you can fund this with non-dilutive capital or not is ultimately determining which side of the, you know, regulatory capture argument you're on. >> Yeah, so regulatory capture I agree on the regulatory capture, but maybe at a different angle. I view it as those leading the F1 race, Anthropic, OpenAI, you don't want others to catch up to you. The way that that could ultimately be called throwing sand in the gears. >> you think that they're leading? Like >> I don't even think there's a question
[5:36] that they're I think I disagree. >> All right, go ahead. >> Good, good. It's good to disagree >> All right, I want to hear about this. >> When it comes to pure models and what they're doing on the enterprise, I think Anthropic and OpenAI have had the clear lead. Do I believe that there's a gap narrowing versus Meta versus Google versus others? Yeah, because But you see, I view it as the value chain is going to be in the data. The models over time will become more commoditized. So,
[6:06] I don't become as focused on open-source models as happen China. What Muse has done is actually just a positive for overall AI. So, I view it differently rather than like the recent capital component, I view it more like you're way ahead in the race. Others are starting to catch up. I could argue like the Dario essay, that was 20% of that added to Meta's stock. Because Zuck looks at that and be like, "Okay, you're going to go right lane 55 miles, minivan, with a bumper sticker saying
[6:36] saying your kid was an honor student in third grade. I'll go in the left lane, Ferrari, and we'll gain share." And that's essentially what Zuck's done. >> So, I think that the private large lab models are in way more trouble than people realize. And I'll give you two uh maybe examples. First is my belief is that token consumption leads revenue. So, if token consumption starts ramping aggressively, revenue follows. I believe that the revenue has been
[7:07] plateauing inside of these large language models, Anthropic in particular, mainly because the general intelligence approach is not nearly as valuable or large as people originally thought. Because I think the original thought process was you're going to create the smartest model possible, and then everyone is going to use it for everything. Now, what's happening though is people are saying, "Wait a second, I want to own my own intelligence." >> Sa- Sa- sovereign AI. >> I want to have applied AI for this specific use case, and maybe I use them to get started, because it's easy for me
[7:38] to build some software, plug them in. But if you take a Look at what we did with Silvia. We started off 100% as a chat GPT wrapper. Over time we said well rather than use chat GPT let's switch over to Claude which seems to be better. Our monthly compute expense went from zero to hundreds of thousands of dollars. We were one of the fastest month over month growing customers I think that Anthropic had. Great, they loved us. Then we said wait a minute, why are we using from an architecture standpoint model to do certain things? So there was literally a point where you would
[8:08] refresh the page on Silvia and you'd hit the model. We're like let's stop doing that, right? So you start to get smarter about like when do we hit the model? How do we hit the model? But then over time it eventually became why don't we just train our own models? Why don't we build our own harness? Why don't we go and actually have the sovereignty so we're not having to use somebody else's intelligence. We can also though make it custom. So now we get better performance, we get better accuracy, we get lower cost and we get less latency. So we're not going to give up on using them. We'll still use them for plenty of
[8:38] stuff. But the token consumption that they have from us versus what we're doing ourselves is shifting and so on like a per customer basis it seems like a lot of companies are saying hey I'm not going to go to zero with them. But actually I'm going to use them less and less as I build my own stuff and to me that feels like a massive headwind that not a lot of people are talking about. >> So I think on some early adopt new adopt look sovereign AI and Karp talks about it Jensen talks about it as well. Like that that's the golden goose. It's about
[9:09] companies now they've gotten their arms around okay that data we ultimately want to maintain and we don't want models coming into that data to replicate our business and ultimately put us out of business. I think the difference is that when you look look 4% of companies have gone down the AI path just in the US today. >> Only 4%? >> 4%. >> Wow. >> So that that's the difference on that. So in other words if we were at 50% and that can some I'd
[9:39] >> So you think there's a lot of runway to go. We're still in the zero to one phase of just AI adoption in general. It's only the people who are on kind of the the frontier. They're the ones who are starting to adopt sovereign AI, but that might be 0.1% penetration. >> No, and I and that's why so many CIOs that I talked to, okay, the ones that are very advanced, exactly where you are and that's where like a sovereign AI, you talk about like what Palantir, Nvidia, some of the things even, you know, I think some of the infrastructure players are looking at. But when you think like look, Europe, they're building Blockbuster videos.
[10:11] You know, you think about what's happening India, they're trying to figure out do they go US China, Middle East very skewed toward US tech in terms of what we've seen there. So it's a two-horse race between the US and China. When you think about why we're so bullish, it's about the numbers. Europe, eventually they're going to, you know, once through the red tape, they'll ultimately embrace AI. Middle East, India, and then think about you're going to get to another 60-70% of enterprises that go after AI and adopt. That's still
[10:41] on the come. And then think about on the consumer side. That's just starting now with Apple finally actually having AI. You see what Meta's doing as well in terms of Muse. So my view is like we still have another two, three years of just massive growth. And I think that's something that investors are still underestimating the scale and scope of. >> Now, when you see uh all these leaders come together in Washington D.C. They meet with the president. Uh whether
[11:11] people like him or not, one thing I think everyone should agree on is that he's good at branding. And he says his artificial intelligence name sucks. I want to name it super intelligence. What are you going to call it? >> Look, I mean, we still call it AI. So I'm not But the point is like forget the name thing. Trump basically had to get everyone in a room and be like, "Look, but like what's going on?" And because you go into a midterm election cycle, if you continue to be dystopian, talk negative, you know, about like what
[11:41] this could do to mankind, guess what? More data centers get shut down. It's becoming a huge I think it's becoming a huge debate no matter what the town, what the state, what the city is. It's very important that you get everyone on the same page because every data center that gets shut down, China wins. And that is I think the most important part of where we are. You need more self-regulation rather than public-private beltway talking points,
[12:13] grandstanding type stuff because I think that would be the most negative thing. My whole view is like for someone like myself who's covered tech my whole career, 3 and 1/2 million air miles, so many times like in Taiwan, in Korea, you see them building fabs 18 hours a day with one bathroom break. Then I land in Newark Airport, there's a fistfight at Dunkin' Donuts, and you wonder why we're 17th in math. Now for the first time, US is ahead of China when it comes to tech. You don't
[12:43] want that to end. >> Today's episode is brought to you by Arch Public. Arch Public has just expanded its agentic trading platform beyond crypto, so pay attention. This is a big one. Now they are automating strategies across stocks, commodities, and ETFs, and I think that this is going to be huge. You can now automatically take profits when one market hits new all-time highs and rotate that capital into other markets showing more opportunity. Whether you're rotating capital into AI stocks, gold, if you're investing in the S&P 500, or you're accumulating Bitcoin, Arch Public brings real discipline and automation to your
[13:14] investment strategy. Additionally, they've launched a powerful new tax loss harvesting tool. With crypto being so volatile and its exemption from the wash sale rule, Arch Public can offset gains with losses without compromising your long-term positions. It's exactly what every serious investor does. Institutional grade automation that works across every major asset class. There's no more emotional trading, no more missing tax opportunities, just smarter, hands-free execution of your preferred strategies. Go to archpublic.com right now. Connect with their team, set
[13:44] up a time, bring your account, and if you'd like, then you can learn what automated trading can do for you. archpublic.com >> Do you worry about America pursuing a closed source model versus China pursuing a open weight open source model, or do you think that actually American open source needs to become much more prevalent? >> think that's what Nvidia has done. I mean, you go out to the hugging acquisition. I mean, that cuz Jensen understands better than anyone. Forget it, like >> He He does feel like the biggest adult
[14:16] in the room by far. >> It's not even a question. He is the adult in the I actually think the two that understand it the most are Jensen and Karp. >> Interesting. >> Because they understand just forget the model cuz everyone focus on the models and they have a much more, I think, global view because of I think in terms of Palantir where they ultimately work with in terms of enterprise, but specifically government. I think Jensen has the most global view. Jensen understands open
[14:46] source you need just based on cost. Think about it. Let's say phones, going back 2007, 2008. Let's say Apple priced the phones double where they were in 2007, 2008. Would you have 1.5 billion iPhones today? No. You have to make sure a tokenization cost, the pricing, open source. That's why I think Muse also, just as a as a side note, is very important. And that's why, like, I think the most important
[15:16] person that has the best perspective is Nvidia, and he continues to sort of be the calming force. Even when you did the all-in podcast, you know, right after the Dario essay sub opera on the weekend, and that was like such an important time for the market. >> Well, so this goes back to these large language models, right? I think everyone now has seen this chart where all of the token prices they're pushing out on the frontier in terms of their intelligence, but they're coming down in price. And so you have this essentially a price war occurring. So, if you go back to well, what is the revenue of a model app? It
[15:46] is token price times token consumption equals revenue. So, if you have consumption per customer dropping, and you have token price dropping, it just feels like there's massive pressure. So, then the question becomes Anthropic was supposed to go public this year. OpenAI was and then came out and said they're not going to. Do you think we get the Anthropic IPO before the end of the year? >> Yeah, I think. I mean, I think that's like I'd be shocked if that doesn't happen. >> Mhm. >> And then >> Well, we've seen uh take Oura Ring. Those guys, they claim they were 4x oversubscribed. They
[16:17] claim, you know, all of these great metrics going into the IPO, and then the night of uh the night before they basically pull it. I don't know what happened there, right? But it does feel like and they blamed, by the way, market conditions, which is kind of like a nonsense thing, right? But fine. >> I couldn't get to the dinner cuz of traffic. >> Yeah. Yeah. Yeah. >> We should definitely do it again. >> [laughter] >> So, uh do you feel like Anthropic, same thing, is like maybe it's not an Anthropic problem, but they just say look at the market and say there's too much here going into a midterm, let's not let's not go out. >> I just think after the mid after the
[16:48] midterm it happen. Look, to me, it's one First, I think also like when it comes like public, it's always hard to compare cuz be like this IPO got pushed or that Anthropic is a is a whole different game. Just because globally, and I think we saw this with SpaceX as well, the appetite to understand what the trajectory looks like, they'll look do they have unique risk factors? Like mankind, you You of course. But I do think we're going into a golden
[17:19] age for AI. We're going into a golden age where many of the private companies ultimately will be public. And I think that's healthy. Because for the overall tech trade, Anthropic and OpenAI being public is good. It's not bad. I think it actually and I think that's ultimately where it's going. SpaceX, OpenAI, Anthropic being public are are actually very good for the market. >> I agree. I just think that when I talk with people, there is this blind belief, these are amazing
[17:49] companies. And when I push on, specifically people who aren't like analyst, why do you think they're a good company? They start talking about the quality of the product. And maybe my word of caution to people is just like a good product does not equal a good company. You can have a good product and a good company. One It is undeniable the products are amazing. People love using them. But when you start to hear some of the data points that are coming out, you know, in terms of what's in the uh the S-1, you look at some of the the uh kind of trends, I just think there's way more headwind than uh than maybe people think because
[18:20] you're dealing with a you're dealing with an industry where look at Jev, right? Maybe as as a good example. Like that wasn't on the chessboard 3 weeks ago. >> Mhm. >> All of a sudden, zero to $100 million run rate in you know, the first week and scaling. And I use our engineering team as a uh maybe a a pressure point for me. So, I go ask them, I say, "Hey, how many of our queries are routing through various models?" All of a sudden, Jev now, it's not 50% but they're starting to use it. And
[18:50] they're starting to put it in places where previously you couldn't use one of the frontier models cuz it was cost prohibitive. And so, you start to almost like it's like death by 1,000 cuts it feels like. And we're going to get this huge fragmentation, which maybe to your point, that's why they're trying to pull the ladder up, is because that would solidify them as the centralized winner. >> Well, also, Anthropic and OpenAI, they've been very intelligent and smart seeing around the corner, building up the enterprise sales force. >> Mhm. >> Going after come cuz they recognize that
[19:21] been pretty early on like the consumer piece is not where they're going to monetize. They're going to monetize on the enterprise. But then it goes back to like listen 5% of companies have gone down the AI path. Take a company just given the amount of data they ingest, they start to go to AI, what that's going to do from an incremental business perspective for Anthropic, for open AI, but for the overall industry, it speaks to >> Well, even meta, right? News for small business. They're all, you know, going that way. >> And that's a good example like with meta like everyone counted them out. They're
[19:52] spending like 19 years rock stars. You're not seeing the actual, you know, return. When you have 3 and 1/2 billion users, you could be late. They look at it now what time of the stock. Apple same thing, right? Apple bear set stocks it should be like 200 to fit They don't get it because it's about the install base. 1.5 billion iPhones, 2.5 billion iOS devices. You're now a toll collector on the consumer AI highway. I just think whether you're bullish on Anthropic or bearish or open
[20:24] AI take a step back. This AI trade is spreading. >> Mhm. >> Not just from memory to infrastructure to a lot of even the traditional players, the HPs, the Dells, the Cisco it's going to go to energy. It's going to go to infrastructure. And I think that is a very key theme and I just think a lot of investors that you know, have missed really every transformational growth stock the last 25 years it's very I think it's a
[20:54] it's a part of the market where you have to be able to look out 2 3 years. Physical AI globally what I believe is it going to be US tech dominance and what that means for innovation this country, and I just continue also believe the taking jobs away narrative, just like the SaaS apocalypse, I think was way, way overdone. And I think now you're starting to see some of that play out. More jobs will be created
[21:24] by AI than taken away over the next decade in my view. >> Easy. And look, we see this inside of our company, there are more agents than humans. There are jobs that we did not go hire for because the agents can do it. But, because the agents are so productive, the company is growing faster than it would otherwise, which then leads to us being able to hire more people. >> And innovation in this country, it used to be like, if you wanted to do a startup, you got to get seed funding this 20 million, 30. Now, because of the
[21:55] engineering horsepower that you could do, think about the innovation that now could happen in this country. And I just think it's a it goes back to like the White House meeting and big tech and where we are. Everyone recognized like this is a moment in time. And it is just very important that you do not put the brakes on. You put the brakes on, we will look back 5, 10 years from now be like, how did China get so ahead of us? And I think it's just it's a very, very
[22:25] sensitive, like whatever you call it, like hot button topic, but I don't really see any gray in terms of the direction that we have to go. >> So, let's say I'm an investor. I'm looking at the entire public markets, and I'm saying to myself, all right, there's a lot here to unpack. I've got the mag seven, I've got the infrastructure, I've got all the data centers, I've got memory, I've got, you know, pure play AI. How do I think through allocating money? Right? Is it put a certain percentage of my portfolio in AI overall, or these like different
[22:56] sectors, but like the entire economy now is AI, and so of course you've got to, you know, have it throughout your portfolio? Just like, how are you talking to me about this? >> It's about second, third, fourth derivatives. Like I I continue to believe large cap tech is starting like flex its muscles again. You look like Nvidia is a good example. Like those that continue to sort of you know, defy or that they think it's getting tired, it's not a shiny new object anymore. You could argue that that stock is way cheaper today than it was 6 to 12 months ago, and you still haven't even gone into what's going to
[23:27] be the full wave of physical AI and and where we're playing. Our Asia check showed 13 to 1 demand. I demand and supply 13 to 1 for chips in Asia. That's also why like AMD. Count if you look like what Lisa Su's done, they're just starting in terms on the chip side. I just think you have to own chips. You got to own the hyperscalers, and I think you've seen specifically Microsoft, Amazon, Google like reinserting themselves because of the install base, and it goes back to like less than 5% that have gone down the AI
[23:59] path as more head down that path. That's bullish for the hyperscalers. The software, the SaaS apocalypse, I think it was a fictional narrative, and I think there's a lot of these if you look like from Salesforce to ServiceNow. Of course, Palantir weeding in, and I think that's one that's almost going to become a category changer, especially as the free cash flow ramps significantly. Cybersecurity, another example. Go back to like April. Cybersecurity is done. Anthropic's eating their lunch. Look at CrowdStrike, Palo Alto, Zscaler. Look at those
[24:29] stocks. Now look at it. And then I think you have to even go further out and look like infrastructure and energy. My whole point is from an allocation perspective, you own the core names and winners and themes, but then you have to see where everything is heading. Like when when when like look at Tesla. Like when are they actually going to start to now transition from that EV player to ultimately robotaxis, autonomous, optimist. Once the transition happened,
[25:01] you've already missed it. >> Mhm. Personal agents are all the rage now. You've got, you know, the Grok boxes of the world who are kind of sort of personal agents, but then you've got Instinct, Muse, etc. How does that impact the AI market? And [clears throat] those are consumer. It seems like these businesses, at least Instinct and Zoc have both come out and said that they are not going to try to charge the consumer. They're going to find other ways to monetize. >> Yeah. >> What do you just think about that as a sector and and how investors should think about it? >> that's going to be like a whole 'nother area of investing. Because my whole
[25:32] point is like you have to think about like the enterprise, then the consumer piece. Clearly like the infrastructure, if it was the gold rush, there was the picks and shovels, like they're ultimately going to benefit on either side. But if you think about on the consumer, we're just now starting the consumer ad revolution. That's why I think Apple, the stock's doing what it's doing. And I think you start to see what I believe is going to almost be like a new category that comes on the line. We're going to be talking about stocks
[26:02] 18 24 months from now that we we don't we're not talking about today. >> Mhm. >> But that's why it goes back to my view like you don't have equilibrium till late 2028, early 2029 in terms of chips. And I think that's something despite like 10-year oil, geopolitical, whatever the moment and the sort of, you know, the nervousness of the time, I think tech is going to continue to plow through and be the leader of this market.
[26:32] >> interesting because the macro environment, almost everything is telling you, "Hey, top or headwinds, right?" You have interest rates going up, you have oil, you know, 100 bucks or so, you've got 10-year over 5%, uh people are very nervous, uh home prices are, you know, quote-unquote falling. All this stuff that if I wanted to paint you a very >> K-shaped economy. >> Yeah, well, if I wanted to paint you a just a very negative story, I could do it easily. >> But if I told you 6 months ago, we're sitting here I know a new studio and that, but like before the new studio, and I said 6 months from now, oil 105 110,
[27:03] 10-year 5.2 5.3, Iran still going as I told you everything and I said, "Okay, where do you think S&P and Nasdaq are?" You probably would be off by 15% lower. >> Mhm. I agree with that. >> Because it comes down the market is telling you we're in a fourth industrial revolution. The growth, I could argue the 10-year is there because of the growth and because we're going to be ultimately more and more debt raises and because of the data
[27:35] center build out and because of what we're seeing it from a CapEx perspective. So, I just think investors, you can my whole career going back to the '80s, if I got caught up in like geopolitical, Europe's going bankrupt, they Dude, about my mind is like I my career would not be where it is today. So, I'm just saying it's very easy to get caught up in the bears, the catastrophists that have called 10 of the last two downturns to scare retail. And I think
[28:06] that always creates the opportunities. >> Mhm. >> This episode is brought to you by Lava. The Lava Card is the first credit card that allows you to earn up to 5% back in Bitcoin on every single purchase. Most cards have steep annual fees and unclear points programs, but the Lava Card offers a much better deal. Earn real Bitcoin every time that you spend. Plus, there's no annual fee. There's zero FX fees and zero spread when you spend internationally. On top of that, Lava gives you access to a Bitcoin line of credit, yield on cash, instant stablecoin deposits and withdrawals, on
[28:38] and off-ramps to and from your bank account, and much, much more. With Lava, you can borrow against your Bitcoin at the lowest rates, earn yield, move money globally, and stack Bitcoin with each and every purchase you make. Most financial services squeeze you with high fees, confusing rewards programs. Lava keeps it simple and they help you build your long-term savings via your everyday spend. Visit lava.xyz/pomp to get started in just a few minutes today. lava.xyz/pomp to get started today.
[29:09] 99% of Warren Buffett's wealth was accumulated after his 50th birthday. >> [music] >> The asset that made that possible was not in his portfolio though. It was his health. The first rule of investing is you actually got to be around to compound. Investor Health is a platform that we recently created and I want you to go test it to tell us what you think. Investor Health is for people who take their body as seriously as their portfolio. Weight, [music] metabolism, energy, longevity, GLP-1s, NAD+, B12, all prescribed by licensed clinicians.
[29:40] Here's how it works. You complete a short online form. A board-certified clinician reviews it and decides what's right for you. Your medication is compounded and shipped to your door and most patients in 48 to 72 hours receive it. No office visits, no referrals, no waiting rooms. Programs start at $149 a month. It's about $5 a day and it's HSA and FSA eligible. No contracts, [music] cancel anytime and it's available in all 50 states. Go to investorhealth.com to
[30:11] learn more and go long on life. You got to be around to compound. investorhealth.com [music] Investor Health is a telehealth platform, not a medical provider. Medical services are provided by affiliated licensed clinicians. Compounded medications are not FDA approved. >> [music] >> Individual results may vary and treatment requires evaluation by a licensed provider. Terms and conditions apply. See investorhealth.com [music] You recently priced a brand new financial product uh in the market.
[30:42] It is a special purpose investment fund, which is kind of like a SPAC, but it's an investment fund. $200 million raised. And my understanding is that you're going to take this, which is called the AIV Ultra Fund. You're going to the ticker is IVAI. You're going to take it and go invest in the private market. So, it's a publicly traded permanent capital vehicle that will have exposure to private companies. What are you going to look for to to allocate capital? >> And it's the first public company ever to invest in private AI tech companies.
[31:14] Look, for me, so many people around the world, whether it's retail, financial advisors, well, whoever it may be, someone in an airport, how do you invest in private tech companies? Institutions is the big issue, right? There's SPDs, SPV2, SPV3. >> Boy, stop at three. >> [laughter] >> You can go to My whole view is to create a vehicle that you could own the leading private tech companies and give, instead
[31:44] of just a small club outside Silicon Valley that can own them. Why can't people in the public own these private tech companies? And that's what we're doing. I mean, it's going to, you know, this is really kind of a Swiss Army knife approach to try and identify what that portfolio looks like, whether it's some of them are bigger players or well-known today to some that like, who could be the next Palantir? >> Mhm. >> Who could be the next model that that ultimately drives this market? And that was really the origin doing it
[32:14] with partners Ed and Jeff Leather. It Look, it was super exciting to do that because I just believe it's a new frontier. >> Mhm. I think it's a very good idea. Um, and also I think that you are the right person to go and find these private AI companies. And the blending of the public and private markets continues, right? Whether it's tokenization, whether it's funds like this. Like we have seen this play out, and this just feels like another step in that direction. >> Well, it's like why if someone's like you're in a private credit fund, why can you be in a private tech product
[32:46] where you own private AI tech? Look, it comes down to like this is the new landscape. The reality is trillions of dollars are going to be spent in this sector. A lot of it's going to be spent by private company on private companies that are leading it. And trying to identify that and I think if you go back historically, it was a small group of people or companies that could buy these comp- that could buy these private
[33:16] companies. And I think it's just important where we are today to create a public vehicle that does that. >> Yeah. It it's very very interesting also to see the private companies how fast they're growing. Like one of the companies I'm an investor in is called MicroOne. And MicroOne is a data business, but the data business is, you know, Scale AI's like a data labeling business. These guys have some data labeling to them, but also they are going out and they're paying doctors and lawyers, etc. to go and actually create data for them to then go and sell through. They're now going to companies
[33:47] and they'll pay pretty large amounts of it. Um and I'll give you one that I'd never heard of before. Somebody emailed me this week and said, "Hey, will you license us the podcast?" Because they want the data. I don't know if we're going to do it. I have no clue what that even means. We haven't talked to them yet, whatever, but I was just like, "Oh, data is data." Like there's going to be data all over the place that people are going to try to buy up and use for all these different >> Data is the new oil and gold. Like why do you think Amazon put the gates up with Muse?
[34:18] Because the reality is that and that's why sovereign AI is such an important dynamic, which you could say is negative for the mono players, cuz companies say, "Okay, we want to have control of our data. We're going to do it in our walls and I think that's See, different than the model players that are narrowing the gap, I actually think sovereign AI is more of the bigger debate. >> Why? >> Because
[34:48] companies are starting to get their sea legs. They're starting to understand like, okay, we got to go full AI. We need these strategic paths, but is it just giving it an out sourcing it to and then I think >> like there's no way that's going to happen. >> So that's So that's So it could you go through evolutions, right? Like it's no different than any technology. The technology it's going to take time for these companies to recognize like trial and error, what's the best path.
[35:19] And I just think like that's going to create so many more opportunities for other companies to monetize. >> I don't think people understand the sovereign AI thing, right? It's own your own intelligence sovereign AI, however you want to call it. Again, I just know the thing that we have seen. If we route a model or a query to our models on our hardware versus go to the frontier models, the cost savings is about 97%. But the cost is not the reason to do it exclusively.
[35:49] It's actually the answers are more accurate. So what I always go back to and and to be honest, the first when we first started going down this path, we had some very smart AI engineers and they were explaining to me what they thought was possible and all that stuff. And you kind of have in the back of your head, dude, there's no way a team of six, seven, eight engineers is going to beat a trillion dollar company. And then they explained to me, well, let me show you how this works. And so today we've come out and we've showed Silvia is more accurate on tax, mortgage, credit cards, you know, these personal finance topics.
[36:19] But it is less about how many engineers you have or anything. It's are you general or are you, you know, applied AI? And then what is the data set that you're able to use to get to that level of accuracy, the the reinforcement learning, all of the eval rubrics, like all of these components. And I just started to realize I was like, if we can pull this off, dude, what is it going to happen when these large companies start realizing, why don't we just do it ourselves? And they have hundreds, thousands of engineers. Like this is going to completely change the way that people
[36:49] think about this technology being integrated into these companies. >> But is it democratization of the technology? See, I I don't view it as a See, I view it as >> It's empowering for the company. >> This is not just a duopoly. >> Mhm. >> If it was a Let's just say this was just OpenAI and Anthropic. And Nvidia was selling the chips. Like that as a bull case in this market, that's not You need democratiz- You need data sovereignty. And I actually think like
[37:21] this is actually like, like I said, it's third inning of this game. Because it's not Those that say, "Oh, it's sixth, seventh inning." I mean, I go around the country so much like looking at the data centers and I just think even if 10 to 15% get voted down, like you're going to have 800,000 data centers built in the next 12, 18 months. Like that is that's building the hearts and lungs of what's going to be really a new economy that's building out
[37:51] in this in this country. >> It's 100% a new economy. And And I think that I'll leave you with this. I recently was talking with some of the engineers and and I said to them I said, "Would it be offensive if I told you that I'm I've now retrained myself over the last 6 months and I'm a software engineer?" My entire career I've been, you know, I was product manager at Facebook and like understand product. I would never writing code, right? Even if, you know, I was building a WordPress site. It was kind of like you're hacking it together. I'm not a software engineer. I don't have a computer science degree, all this kind of stuff. And for people who are non-technical who work with engineers,
[38:22] there's kind of like a bright line. >> Yes. >> You you you do not claim to do what they do because they are very special. They They're the stars, you're the support staff, right? It's kind of the way I've always thought about it. But I said to him, I said, "Look, here's [snorts] how I have started to think about this." And And one of the engineers, right, one of our best engineers said to me, "If I was doing an interview to hire a software engineer, one of the first questions I ask is show me what you've built." And he's like, "And you have all this stuff that you've built." So, he's like, "The lines are blurring." Now, I'm not going to go actually put software engineer on my Twitter bio, right? You
[38:52] know, like bang my chest and be like, "I'm a software engineer now." But, if that starts to become a skill that dummy me can use, what happens when that permeates throughout society and throughout the economy and now all of a sudden productivity accelerates? >> But that's the whole the whole point that the Fed's talked about is that >> Yeah. >> new technology the last 100 years has been a net job detractor or it's about productivity enhanced. And I just think sometimes as you're going through the
[39:23] storm, the turbulence, you can't see the other side. Sometimes I think it's very easy to get caught up in narratives in this market. And I think that's some of the dangers as an investor, too, when you get too caught up because you cuz I see it from an end user. I see it from CIO. I see it in Asia. I see it from a chip brick. When you see the innovate, you see physical AI that's ultimately coming. And that's why like when you have adults in the room like Jensen kind of navigating it, I think that you just
[39:54] continue sort That's the yellow brick road to the success. >> Yeah, it's fascinating. Where can someone go to find you online? >> Yeah, so dives tech. And again, I don't block hardly anyone or or or LinkedIn. And No, look, I appreciate Look, this is just it's an important time to sort of, you know, navigate the narratives in some of these white-knuckle moments and make sure investors just don't get, you know, sort of caught up sometimes in the in the in these battles from the hibernation mode.
[40:24] >> I think you're doing a great job. I'm very excited about >> No, thank you so much. >> IVAI is is the ticker and we'll do this again. I'm going to convince you to start doing this weekly. >> Hey, I'm I'm here for it. >> [laughter] >> All right, we'll see everyone soon. >> Thanks.
[0:31] Yorkville Ives, and he's here to talk all about the AI trade. We talk about the sectors that he's interested in, the sectors where he thinks value's going to accrue, and maybe some areas where you should be cautious. We talk about the private AI model labs, where is pressure coming from, or is AI actually going to kill us all, what happened at the White House meeting, and then we even get into what's going on with all of the personal assistants that have been coming to the market. Dan's got some very unique thoughts on the AI trade, and I think that if you watch this, it's going to make you a smarter investor in the market. And then Dan finally finishes up, and he talks about the Ives Ultra Fund. It's a brand new fund that he's
[1:01] brought to market. It trades under the ticker IVAI, and I think people are going to be very interested to see how they can hold a public stock and get exposure to the leading private AI companies. Here's my latest conversation with Dan Ives. All right, Dan. AI trade, obviously, was super hot earlier this year, then got a little cooling off in the summer, but it feels like it's starting to come back here a little bit. And we have the private companies that are knocking on the door of going public in Anthropic, Open AI. But the public stocks, where do you see value through the end of the year? How how are you thinking about the AI trade today?
[1:31] >> Yeah. Look, I think the demand, I think it started last quarter. That's where it changed. Cuz the monetization you saw specifically with Microsoft and the and the other hyperscalers, I think that was an inflection point quarter. >> [snorts] >> I think now a lot of it is on the second, third, fourth derivatives. And then, of course, memory stocks were there. But I think you're seeing it across infrastructure with Dell, with Cisco. I think with cybersecurity stocks, you're starting to see it play out. Look, take a step back. We're less than 15% through what the spending
[2:03] trend's going to be the next 3 to 4 years. It's 4 to 5 trillion dollars. For every dollar of CapEx, there's a $5 to $6 multiplier across the rest of tech. So, I think we're going to go through these sort of ebbs and flows, the regulatory capture, the Dalio essay. I mean, some of these things will continue to sort of get in the way to some extent in the near term of the trade, but I just continue to believe we are in the early days of what's going to be a multi-year tech bull market. The first time in 30 years, US has added China when it comes to tech.
[2:33] >> When you see Dalio come out and basically say, "Hey, we think that there's a 5% chance or 10% chance of AI killing us all. We must slow down." And, you know, kind of all of these calls for essentially like, "Hey, government, regulate me harder." I've never heard a business that is serious talk this way. What's your general read? Are they sincere in this? Are they just trying to use it to slow down their competitors so they can catch up? Or or what's going on? >> I mean, 1910, the guy who said, "We can't go down this car route. We have to stay horse and buggy. These cars kill
[3:03] you." I'm I'm just saying, the it's technology. Technology always goes ahead of regulatory. And, I think some of it is I'll call regulatory capture. Like, you go to the top floor and then you pull the ladder up. Mhm. So, I think there is some of that going on. And, I think that's also why you had to have that sort of all hands on meeting, you know, with Trump at at the White House because you have to get everyone on the same page. Cuz the reality is going into a midterm election cycle, you know, >> [clears throat] >> talk like that, and me and you have talked about this before. I think a lot
[3:33] of the tech industry, they created the PR nightmare, the black eye that now they're contending with. But, it's it's an arms race. And, if we slow down, China accelerates. That's not a debate. So, that's why every time I in terms of like the politicians in the Beltway, many, you know, do you want people that are still using Blackberries and flip phones to ultimately be the ones that are, you know, the direction of innovation in this country? And I think that Look, it's not going to slow down. And I think you saw how stocks reacted.
[4:06] I think if you go back I go over the weekend the bears that have called 10 of the last two downturns. They're like, "Oh, my cuz remember the bulls are usually watching football hanging out with their family." And the bears on Saturday, that's usually where it peaks out. But then the reality is is that you you saw what Jensen talked about, what Nadella talked about, what Karp talks about in terms of sovereign AI. That's more what I view as like a very sort of middle-of-the-road view, and that's my opinion. And this is not slowing down anytime soon.
[4:36] >> What I found interesting was, you know, right after the call for the slowdown or pacing or whatever the hell they were saying, um Zuck came out and Zuck did not talk about slowing down. Zuck talked about, "Hey, it's our job to release products that are safe." And my general framework, but I want to hear if you agree with this or not, is the companies that are losing money, that have negative cash flow, they're the ones calling for the slowdown. The companies that are profitable like the Facebooks, the Googles, etc., I don't hear them calling for a slowdown at all. And it almost feels like depending on whether
[5:06] you can fund this with non-dilutive capital or not is ultimately determining which side of the, you know, regulatory capture argument you're on. >> Yeah, so regulatory capture I agree on the regulatory capture, but maybe at a different angle. I view it as those leading the F1 race, Anthropic, OpenAI, you don't want others to catch up to you. The way that that could ultimately be called throwing sand in the gears. >> you think that they're leading? Like >> I don't even think there's a question
[5:36] that they're I think I disagree. >> All right, go ahead. >> Good, good. It's good to disagree >> All right, I want to hear about this. >> When it comes to pure models and what they're doing on the enterprise, I think Anthropic and OpenAI have had the clear lead. Do I believe that there's a gap narrowing versus Meta versus Google versus others? Yeah, because But you see, I view it as the value chain is going to be in the data. The models over time will become more commoditized. So,
[6:06] I don't become as focused on open-source models as happen China. What Muse has done is actually just a positive for overall AI. So, I view it differently rather than like the recent capital component, I view it more like you're way ahead in the race. Others are starting to catch up. I could argue like the Dario essay, that was 20% of that added to Meta's stock. Because Zuck looks at that and be like, "Okay, you're going to go right lane 55 miles, minivan, with a bumper sticker saying
[6:36] saying your kid was an honor student in third grade. I'll go in the left lane, Ferrari, and we'll gain share." And that's essentially what Zuck's done. >> So, I think that the private large lab models are in way more trouble than people realize. And I'll give you two uh maybe examples. First is my belief is that token consumption leads revenue. So, if token consumption starts ramping aggressively, revenue follows. I believe that the revenue has been
[7:07] plateauing inside of these large language models, Anthropic in particular, mainly because the general intelligence approach is not nearly as valuable or large as people originally thought. Because I think the original thought process was you're going to create the smartest model possible, and then everyone is going to use it for everything. Now, what's happening though is people are saying, "Wait a second, I want to own my own intelligence." >> Sa- Sa- sovereign AI. >> I want to have applied AI for this specific use case, and maybe I use them to get started, because it's easy for me
[7:38] to build some software, plug them in. But if you take a Look at what we did with Silvia. We started off 100% as a chat GPT wrapper. Over time we said well rather than use chat GPT let's switch over to Claude which seems to be better. Our monthly compute expense went from zero to hundreds of thousands of dollars. We were one of the fastest month over month growing customers I think that Anthropic had. Great, they loved us. Then we said wait a minute, why are we using from an architecture standpoint model to do certain things? So there was literally a point where you would
[8:08] refresh the page on Silvia and you'd hit the model. We're like let's stop doing that, right? So you start to get smarter about like when do we hit the model? How do we hit the model? But then over time it eventually became why don't we just train our own models? Why don't we build our own harness? Why don't we go and actually have the sovereignty so we're not having to use somebody else's intelligence. We can also though make it custom. So now we get better performance, we get better accuracy, we get lower cost and we get less latency. So we're not going to give up on using them. We'll still use them for plenty of
[8:38] stuff. But the token consumption that they have from us versus what we're doing ourselves is shifting and so on like a per customer basis it seems like a lot of companies are saying hey I'm not going to go to zero with them. But actually I'm going to use them less and less as I build my own stuff and to me that feels like a massive headwind that not a lot of people are talking about. >> So I think on some early adopt new adopt look sovereign AI and Karp talks about it Jensen talks about it as well. Like that that's the golden goose. It's about
[9:09] companies now they've gotten their arms around okay that data we ultimately want to maintain and we don't want models coming into that data to replicate our business and ultimately put us out of business. I think the difference is that when you look look 4% of companies have gone down the AI path just in the US today. >> Only 4%? >> 4%. >> Wow. >> So that that's the difference on that. So in other words if we were at 50% and that can some I'd
[9:39] >> So you think there's a lot of runway to go. We're still in the zero to one phase of just AI adoption in general. It's only the people who are on kind of the the frontier. They're the ones who are starting to adopt sovereign AI, but that might be 0.1% penetration. >> No, and I and that's why so many CIOs that I talked to, okay, the ones that are very advanced, exactly where you are and that's where like a sovereign AI, you talk about like what Palantir, Nvidia, some of the things even, you know, I think some of the infrastructure players are looking at. But when you think like look, Europe, they're building Blockbuster videos.
[10:11] You know, you think about what's happening India, they're trying to figure out do they go US China, Middle East very skewed toward US tech in terms of what we've seen there. So it's a two-horse race between the US and China. When you think about why we're so bullish, it's about the numbers. Europe, eventually they're going to, you know, once through the red tape, they'll ultimately embrace AI. Middle East, India, and then think about you're going to get to another 60-70% of enterprises that go after AI and adopt. That's still
[10:41] on the come. And then think about on the consumer side. That's just starting now with Apple finally actually having AI. You see what Meta's doing as well in terms of Muse. So my view is like we still have another two, three years of just massive growth. And I think that's something that investors are still underestimating the scale and scope of. >> Now, when you see uh all these leaders come together in Washington D.C. They meet with the president. Uh whether
[11:11] people like him or not, one thing I think everyone should agree on is that he's good at branding. And he says his artificial intelligence name sucks. I want to name it super intelligence. What are you going to call it? >> Look, I mean, we still call it AI. So I'm not But the point is like forget the name thing. Trump basically had to get everyone in a room and be like, "Look, but like what's going on?" And because you go into a midterm election cycle, if you continue to be dystopian, talk negative, you know, about like what
[11:41] this could do to mankind, guess what? More data centers get shut down. It's becoming a huge I think it's becoming a huge debate no matter what the town, what the state, what the city is. It's very important that you get everyone on the same page because every data center that gets shut down, China wins. And that is I think the most important part of where we are. You need more self-regulation rather than public-private beltway talking points,
[12:13] grandstanding type stuff because I think that would be the most negative thing. My whole view is like for someone like myself who's covered tech my whole career, 3 and 1/2 million air miles, so many times like in Taiwan, in Korea, you see them building fabs 18 hours a day with one bathroom break. Then I land in Newark Airport, there's a fistfight at Dunkin' Donuts, and you wonder why we're 17th in math. Now for the first time, US is ahead of China when it comes to tech. You don't
[12:43] want that to end. >> Today's episode is brought to you by Arch Public. Arch Public has just expanded its agentic trading platform beyond crypto, so pay attention. This is a big one. Now they are automating strategies across stocks, commodities, and ETFs, and I think that this is going to be huge. You can now automatically take profits when one market hits new all-time highs and rotate that capital into other markets showing more opportunity. Whether you're rotating capital into AI stocks, gold, if you're investing in the S&P 500, or you're accumulating Bitcoin, Arch Public brings real discipline and automation to your
[13:14] investment strategy. Additionally, they've launched a powerful new tax loss harvesting tool. With crypto being so volatile and its exemption from the wash sale rule, Arch Public can offset gains with losses without compromising your long-term positions. It's exactly what every serious investor does. Institutional grade automation that works across every major asset class. There's no more emotional trading, no more missing tax opportunities, just smarter, hands-free execution of your preferred strategies. Go to archpublic.com right now. Connect with their team, set
[13:44] up a time, bring your account, and if you'd like, then you can learn what automated trading can do for you. archpublic.com >> Do you worry about America pursuing a closed source model versus China pursuing a open weight open source model, or do you think that actually American open source needs to become much more prevalent? >> think that's what Nvidia has done. I mean, you go out to the hugging acquisition. I mean, that cuz Jensen understands better than anyone. Forget it, like >> He He does feel like the biggest adult
[14:16] in the room by far. >> It's not even a question. He is the adult in the I actually think the two that understand it the most are Jensen and Karp. >> Interesting. >> Because they understand just forget the model cuz everyone focus on the models and they have a much more, I think, global view because of I think in terms of Palantir where they ultimately work with in terms of enterprise, but specifically government. I think Jensen has the most global view. Jensen understands open
[14:46] source you need just based on cost. Think about it. Let's say phones, going back 2007, 2008. Let's say Apple priced the phones double where they were in 2007, 2008. Would you have 1.5 billion iPhones today? No. You have to make sure a tokenization cost, the pricing, open source. That's why I think Muse also, just as a as a side note, is very important. And that's why, like, I think the most important
[15:16] person that has the best perspective is Nvidia, and he continues to sort of be the calming force. Even when you did the all-in podcast, you know, right after the Dario essay sub opera on the weekend, and that was like such an important time for the market. >> Well, so this goes back to these large language models, right? I think everyone now has seen this chart where all of the token prices they're pushing out on the frontier in terms of their intelligence, but they're coming down in price. And so you have this essentially a price war occurring. So, if you go back to well, what is the revenue of a model app? It
[15:46] is token price times token consumption equals revenue. So, if you have consumption per customer dropping, and you have token price dropping, it just feels like there's massive pressure. So, then the question becomes Anthropic was supposed to go public this year. OpenAI was and then came out and said they're not going to. Do you think we get the Anthropic IPO before the end of the year? >> Yeah, I think. I mean, I think that's like I'd be shocked if that doesn't happen. >> Mhm. >> And then >> Well, we've seen uh take Oura Ring. Those guys, they claim they were 4x oversubscribed. They
[16:17] claim, you know, all of these great metrics going into the IPO, and then the night of uh the night before they basically pull it. I don't know what happened there, right? But it does feel like and they blamed, by the way, market conditions, which is kind of like a nonsense thing, right? But fine. >> I couldn't get to the dinner cuz of traffic. >> Yeah. Yeah. Yeah. >> We should definitely do it again. >> [laughter] >> So, uh do you feel like Anthropic, same thing, is like maybe it's not an Anthropic problem, but they just say look at the market and say there's too much here going into a midterm, let's not let's not go out. >> I just think after the mid after the
[16:48] midterm it happen. Look, to me, it's one First, I think also like when it comes like public, it's always hard to compare cuz be like this IPO got pushed or that Anthropic is a is a whole different game. Just because globally, and I think we saw this with SpaceX as well, the appetite to understand what the trajectory looks like, they'll look do they have unique risk factors? Like mankind, you You of course. But I do think we're going into a golden
[17:19] age for AI. We're going into a golden age where many of the private companies ultimately will be public. And I think that's healthy. Because for the overall tech trade, Anthropic and OpenAI being public is good. It's not bad. I think it actually and I think that's ultimately where it's going. SpaceX, OpenAI, Anthropic being public are are actually very good for the market. >> I agree. I just think that when I talk with people, there is this blind belief, these are amazing
[17:49] companies. And when I push on, specifically people who aren't like analyst, why do you think they're a good company? They start talking about the quality of the product. And maybe my word of caution to people is just like a good product does not equal a good company. You can have a good product and a good company. One It is undeniable the products are amazing. People love using them. But when you start to hear some of the data points that are coming out, you know, in terms of what's in the uh the S-1, you look at some of the the uh kind of trends, I just think there's way more headwind than uh than maybe people think because
[18:20] you're dealing with a you're dealing with an industry where look at Jev, right? Maybe as as a good example. Like that wasn't on the chessboard 3 weeks ago. >> Mhm. >> All of a sudden, zero to $100 million run rate in you know, the first week and scaling. And I use our engineering team as a uh maybe a a pressure point for me. So, I go ask them, I say, "Hey, how many of our queries are routing through various models?" All of a sudden, Jev now, it's not 50% but they're starting to use it. And
[18:50] they're starting to put it in places where previously you couldn't use one of the frontier models cuz it was cost prohibitive. And so, you start to almost like it's like death by 1,000 cuts it feels like. And we're going to get this huge fragmentation, which maybe to your point, that's why they're trying to pull the ladder up, is because that would solidify them as the centralized winner. >> Well, also, Anthropic and OpenAI, they've been very intelligent and smart seeing around the corner, building up the enterprise sales force. >> Mhm. >> Going after come cuz they recognize that
[19:21] been pretty early on like the consumer piece is not where they're going to monetize. They're going to monetize on the enterprise. But then it goes back to like listen 5% of companies have gone down the AI path. Take a company just given the amount of data they ingest, they start to go to AI, what that's going to do from an incremental business perspective for Anthropic, for open AI, but for the overall industry, it speaks to >> Well, even meta, right? News for small business. They're all, you know, going that way. >> And that's a good example like with meta like everyone counted them out. They're
[19:52] spending like 19 years rock stars. You're not seeing the actual, you know, return. When you have 3 and 1/2 billion users, you could be late. They look at it now what time of the stock. Apple same thing, right? Apple bear set stocks it should be like 200 to fit They don't get it because it's about the install base. 1.5 billion iPhones, 2.5 billion iOS devices. You're now a toll collector on the consumer AI highway. I just think whether you're bullish on Anthropic or bearish or open
[20:24] AI take a step back. This AI trade is spreading. >> Mhm. >> Not just from memory to infrastructure to a lot of even the traditional players, the HPs, the Dells, the Cisco it's going to go to energy. It's going to go to infrastructure. And I think that is a very key theme and I just think a lot of investors that you know, have missed really every transformational growth stock the last 25 years it's very I think it's a
[20:54] it's a part of the market where you have to be able to look out 2 3 years. Physical AI globally what I believe is it going to be US tech dominance and what that means for innovation this country, and I just continue also believe the taking jobs away narrative, just like the SaaS apocalypse, I think was way, way overdone. And I think now you're starting to see some of that play out. More jobs will be created
[21:24] by AI than taken away over the next decade in my view. >> Easy. And look, we see this inside of our company, there are more agents than humans. There are jobs that we did not go hire for because the agents can do it. But, because the agents are so productive, the company is growing faster than it would otherwise, which then leads to us being able to hire more people. >> And innovation in this country, it used to be like, if you wanted to do a startup, you got to get seed funding this 20 million, 30. Now, because of the
[21:55] engineering horsepower that you could do, think about the innovation that now could happen in this country. And I just think it's a it goes back to like the White House meeting and big tech and where we are. Everyone recognized like this is a moment in time. And it is just very important that you do not put the brakes on. You put the brakes on, we will look back 5, 10 years from now be like, how did China get so ahead of us? And I think it's just it's a very, very
[22:25] sensitive, like whatever you call it, like hot button topic, but I don't really see any gray in terms of the direction that we have to go. >> So, let's say I'm an investor. I'm looking at the entire public markets, and I'm saying to myself, all right, there's a lot here to unpack. I've got the mag seven, I've got the infrastructure, I've got all the data centers, I've got memory, I've got, you know, pure play AI. How do I think through allocating money? Right? Is it put a certain percentage of my portfolio in AI overall, or these like different
[22:56] sectors, but like the entire economy now is AI, and so of course you've got to, you know, have it throughout your portfolio? Just like, how are you talking to me about this? >> It's about second, third, fourth derivatives. Like I I continue to believe large cap tech is starting like flex its muscles again. You look like Nvidia is a good example. Like those that continue to sort of you know, defy or that they think it's getting tired, it's not a shiny new object anymore. You could argue that that stock is way cheaper today than it was 6 to 12 months ago, and you still haven't even gone into what's going to
[23:27] be the full wave of physical AI and and where we're playing. Our Asia check showed 13 to 1 demand. I demand and supply 13 to 1 for chips in Asia. That's also why like AMD. Count if you look like what Lisa Su's done, they're just starting in terms on the chip side. I just think you have to own chips. You got to own the hyperscalers, and I think you've seen specifically Microsoft, Amazon, Google like reinserting themselves because of the install base, and it goes back to like less than 5% that have gone down the AI
[23:59] path as more head down that path. That's bullish for the hyperscalers. The software, the SaaS apocalypse, I think it was a fictional narrative, and I think there's a lot of these if you look like from Salesforce to ServiceNow. Of course, Palantir weeding in, and I think that's one that's almost going to become a category changer, especially as the free cash flow ramps significantly. Cybersecurity, another example. Go back to like April. Cybersecurity is done. Anthropic's eating their lunch. Look at CrowdStrike, Palo Alto, Zscaler. Look at those
[24:29] stocks. Now look at it. And then I think you have to even go further out and look like infrastructure and energy. My whole point is from an allocation perspective, you own the core names and winners and themes, but then you have to see where everything is heading. Like when when when like look at Tesla. Like when are they actually going to start to now transition from that EV player to ultimately robotaxis, autonomous, optimist. Once the transition happened,
[25:01] you've already missed it. >> Mhm. Personal agents are all the rage now. You've got, you know, the Grok boxes of the world who are kind of sort of personal agents, but then you've got Instinct, Muse, etc. How does that impact the AI market? And [clears throat] those are consumer. It seems like these businesses, at least Instinct and Zoc have both come out and said that they are not going to try to charge the consumer. They're going to find other ways to monetize. >> Yeah. >> What do you just think about that as a sector and and how investors should think about it? >> that's going to be like a whole 'nother area of investing. Because my whole
[25:32] point is like you have to think about like the enterprise, then the consumer piece. Clearly like the infrastructure, if it was the gold rush, there was the picks and shovels, like they're ultimately going to benefit on either side. But if you think about on the consumer, we're just now starting the consumer ad revolution. That's why I think Apple, the stock's doing what it's doing. And I think you start to see what I believe is going to almost be like a new category that comes on the line. We're going to be talking about stocks
[26:02] 18 24 months from now that we we don't we're not talking about today. >> Mhm. >> But that's why it goes back to my view like you don't have equilibrium till late 2028, early 2029 in terms of chips. And I think that's something despite like 10-year oil, geopolitical, whatever the moment and the sort of, you know, the nervousness of the time, I think tech is going to continue to plow through and be the leader of this market.
[26:32] >> interesting because the macro environment, almost everything is telling you, "Hey, top or headwinds, right?" You have interest rates going up, you have oil, you know, 100 bucks or so, you've got 10-year over 5%, uh people are very nervous, uh home prices are, you know, quote-unquote falling. All this stuff that if I wanted to paint you a very >> K-shaped economy. >> Yeah, well, if I wanted to paint you a just a very negative story, I could do it easily. >> But if I told you 6 months ago, we're sitting here I know a new studio and that, but like before the new studio, and I said 6 months from now, oil 105 110,
[27:03] 10-year 5.2 5.3, Iran still going as I told you everything and I said, "Okay, where do you think S&P and Nasdaq are?" You probably would be off by 15% lower. >> Mhm. I agree with that. >> Because it comes down the market is telling you we're in a fourth industrial revolution. The growth, I could argue the 10-year is there because of the growth and because we're going to be ultimately more and more debt raises and because of the data
[27:35] center build out and because of what we're seeing it from a CapEx perspective. So, I just think investors, you can my whole career going back to the '80s, if I got caught up in like geopolitical, Europe's going bankrupt, they Dude, about my mind is like I my career would not be where it is today. So, I'm just saying it's very easy to get caught up in the bears, the catastrophists that have called 10 of the last two downturns to scare retail. And I think
[28:06] that always creates the opportunities. >> Mhm. >> This episode is brought to you by Lava. The Lava Card is the first credit card that allows you to earn up to 5% back in Bitcoin on every single purchase. Most cards have steep annual fees and unclear points programs, but the Lava Card offers a much better deal. Earn real Bitcoin every time that you spend. Plus, there's no annual fee. There's zero FX fees and zero spread when you spend internationally. On top of that, Lava gives you access to a Bitcoin line of credit, yield on cash, instant stablecoin deposits and withdrawals, on
[28:38] and off-ramps to and from your bank account, and much, much more. With Lava, you can borrow against your Bitcoin at the lowest rates, earn yield, move money globally, and stack Bitcoin with each and every purchase you make. Most financial services squeeze you with high fees, confusing rewards programs. Lava keeps it simple and they help you build your long-term savings via your everyday spend. Visit lava.xyz/pomp to get started in just a few minutes today. lava.xyz/pomp to get started today.
[29:09] 99% of Warren Buffett's wealth was accumulated after his 50th birthday. >> [music] >> The asset that made that possible was not in his portfolio though. It was his health. The first rule of investing is you actually got to be around to compound. Investor Health is a platform that we recently created and I want you to go test it to tell us what you think. Investor Health is for people who take their body as seriously as their portfolio. Weight, [music] metabolism, energy, longevity, GLP-1s, NAD+, B12, all prescribed by licensed clinicians.
[29:40] Here's how it works. You complete a short online form. A board-certified clinician reviews it and decides what's right for you. Your medication is compounded and shipped to your door and most patients in 48 to 72 hours receive it. No office visits, no referrals, no waiting rooms. Programs start at $149 a month. It's about $5 a day and it's HSA and FSA eligible. No contracts, [music] cancel anytime and it's available in all 50 states. Go to investorhealth.com to
[30:11] learn more and go long on life. You got to be around to compound. investorhealth.com [music] Investor Health is a telehealth platform, not a medical provider. Medical services are provided by affiliated licensed clinicians. Compounded medications are not FDA approved. >> [music] >> Individual results may vary and treatment requires evaluation by a licensed provider. Terms and conditions apply. See investorhealth.com [music] You recently priced a brand new financial product uh in the market.
[30:42] It is a special purpose investment fund, which is kind of like a SPAC, but it's an investment fund. $200 million raised. And my understanding is that you're going to take this, which is called the AIV Ultra Fund. You're going to the ticker is IVAI. You're going to take it and go invest in the private market. So, it's a publicly traded permanent capital vehicle that will have exposure to private companies. What are you going to look for to to allocate capital? >> And it's the first public company ever to invest in private AI tech companies.
[31:14] Look, for me, so many people around the world, whether it's retail, financial advisors, well, whoever it may be, someone in an airport, how do you invest in private tech companies? Institutions is the big issue, right? There's SPDs, SPV2, SPV3. >> Boy, stop at three. >> [laughter] >> You can go to My whole view is to create a vehicle that you could own the leading private tech companies and give, instead
[31:44] of just a small club outside Silicon Valley that can own them. Why can't people in the public own these private tech companies? And that's what we're doing. I mean, it's going to, you know, this is really kind of a Swiss Army knife approach to try and identify what that portfolio looks like, whether it's some of them are bigger players or well-known today to some that like, who could be the next Palantir? >> Mhm. >> Who could be the next model that that ultimately drives this market? And that was really the origin doing it
[32:14] with partners Ed and Jeff Leather. It Look, it was super exciting to do that because I just believe it's a new frontier. >> Mhm. I think it's a very good idea. Um, and also I think that you are the right person to go and find these private AI companies. And the blending of the public and private markets continues, right? Whether it's tokenization, whether it's funds like this. Like we have seen this play out, and this just feels like another step in that direction. >> Well, it's like why if someone's like you're in a private credit fund, why can you be in a private tech product
[32:46] where you own private AI tech? Look, it comes down to like this is the new landscape. The reality is trillions of dollars are going to be spent in this sector. A lot of it's going to be spent by private company on private companies that are leading it. And trying to identify that and I think if you go back historically, it was a small group of people or companies that could buy these comp- that could buy these private
[33:16] companies. And I think it's just important where we are today to create a public vehicle that does that. >> Yeah. It it's very very interesting also to see the private companies how fast they're growing. Like one of the companies I'm an investor in is called MicroOne. And MicroOne is a data business, but the data business is, you know, Scale AI's like a data labeling business. These guys have some data labeling to them, but also they are going out and they're paying doctors and lawyers, etc. to go and actually create data for them to then go and sell through. They're now going to companies
[33:47] and they'll pay pretty large amounts of it. Um and I'll give you one that I'd never heard of before. Somebody emailed me this week and said, "Hey, will you license us the podcast?" Because they want the data. I don't know if we're going to do it. I have no clue what that even means. We haven't talked to them yet, whatever, but I was just like, "Oh, data is data." Like there's going to be data all over the place that people are going to try to buy up and use for all these different >> Data is the new oil and gold. Like why do you think Amazon put the gates up with Muse?
[34:18] Because the reality is that and that's why sovereign AI is such an important dynamic, which you could say is negative for the mono players, cuz companies say, "Okay, we want to have control of our data. We're going to do it in our walls and I think that's See, different than the model players that are narrowing the gap, I actually think sovereign AI is more of the bigger debate. >> Why? >> Because
[34:48] companies are starting to get their sea legs. They're starting to understand like, okay, we got to go full AI. We need these strategic paths, but is it just giving it an out sourcing it to and then I think >> like there's no way that's going to happen. >> So that's So that's So it could you go through evolutions, right? Like it's no different than any technology. The technology it's going to take time for these companies to recognize like trial and error, what's the best path.
[35:19] And I just think like that's going to create so many more opportunities for other companies to monetize. >> I don't think people understand the sovereign AI thing, right? It's own your own intelligence sovereign AI, however you want to call it. Again, I just know the thing that we have seen. If we route a model or a query to our models on our hardware versus go to the frontier models, the cost savings is about 97%. But the cost is not the reason to do it exclusively.
[35:49] It's actually the answers are more accurate. So what I always go back to and and to be honest, the first when we first started going down this path, we had some very smart AI engineers and they were explaining to me what they thought was possible and all that stuff. And you kind of have in the back of your head, dude, there's no way a team of six, seven, eight engineers is going to beat a trillion dollar company. And then they explained to me, well, let me show you how this works. And so today we've come out and we've showed Silvia is more accurate on tax, mortgage, credit cards, you know, these personal finance topics.
[36:19] But it is less about how many engineers you have or anything. It's are you general or are you, you know, applied AI? And then what is the data set that you're able to use to get to that level of accuracy, the the reinforcement learning, all of the eval rubrics, like all of these components. And I just started to realize I was like, if we can pull this off, dude, what is it going to happen when these large companies start realizing, why don't we just do it ourselves? And they have hundreds, thousands of engineers. Like this is going to completely change the way that people
[36:49] think about this technology being integrated into these companies. >> But is it democratization of the technology? See, I I don't view it as a See, I view it as >> It's empowering for the company. >> This is not just a duopoly. >> Mhm. >> If it was a Let's just say this was just OpenAI and Anthropic. And Nvidia was selling the chips. Like that as a bull case in this market, that's not You need democratiz- You need data sovereignty. And I actually think like
[37:21] this is actually like, like I said, it's third inning of this game. Because it's not Those that say, "Oh, it's sixth, seventh inning." I mean, I go around the country so much like looking at the data centers and I just think even if 10 to 15% get voted down, like you're going to have 800,000 data centers built in the next 12, 18 months. Like that is that's building the hearts and lungs of what's going to be really a new economy that's building out
[37:51] in this in this country. >> It's 100% a new economy. And And I think that I'll leave you with this. I recently was talking with some of the engineers and and I said to them I said, "Would it be offensive if I told you that I'm I've now retrained myself over the last 6 months and I'm a software engineer?" My entire career I've been, you know, I was product manager at Facebook and like understand product. I would never writing code, right? Even if, you know, I was building a WordPress site. It was kind of like you're hacking it together. I'm not a software engineer. I don't have a computer science degree, all this kind of stuff. And for people who are non-technical who work with engineers,
[38:22] there's kind of like a bright line. >> Yes. >> You you you do not claim to do what they do because they are very special. They They're the stars, you're the support staff, right? It's kind of the way I've always thought about it. But I said to him, I said, "Look, here's [snorts] how I have started to think about this." And And one of the engineers, right, one of our best engineers said to me, "If I was doing an interview to hire a software engineer, one of the first questions I ask is show me what you've built." And he's like, "And you have all this stuff that you've built." So, he's like, "The lines are blurring." Now, I'm not going to go actually put software engineer on my Twitter bio, right? You
[38:52] know, like bang my chest and be like, "I'm a software engineer now." But, if that starts to become a skill that dummy me can use, what happens when that permeates throughout society and throughout the economy and now all of a sudden productivity accelerates? >> But that's the whole the whole point that the Fed's talked about is that >> Yeah. >> new technology the last 100 years has been a net job detractor or it's about productivity enhanced. And I just think sometimes as you're going through the
[39:23] storm, the turbulence, you can't see the other side. Sometimes I think it's very easy to get caught up in narratives in this market. And I think that's some of the dangers as an investor, too, when you get too caught up because you cuz I see it from an end user. I see it from CIO. I see it in Asia. I see it from a chip brick. When you see the innovate, you see physical AI that's ultimately coming. And that's why like when you have adults in the room like Jensen kind of navigating it, I think that you just
[39:54] continue sort That's the yellow brick road to the success. >> Yeah, it's fascinating. Where can someone go to find you online? >> Yeah, so dives tech. And again, I don't block hardly anyone or or or LinkedIn. And No, look, I appreciate Look, this is just it's an important time to sort of, you know, navigate the narratives in some of these white-knuckle moments and make sure investors just don't get, you know, sort of caught up sometimes in the in the in these battles from the hibernation mode.
[40:24] >> I think you're doing a great job. I'm very excited about >> No, thank you so much. >> IVAI is is the ticker and we'll do this again. I'm going to convince you to start doing this weekly. >> Hey, I'm I'm here for it. >> [laughter] >> All right, we'll see everyone soon. >> Thanks.