Transcription
Greetings everyone. This is Kathy Wood, CEO and CIO of Arc Invest. It is employment Friday. Uh, but there's no employment report. Uh, we will have that next week. But given all of the fireworks uh in the market and all the drama recently, uh I felt it was important uh to to do in the no or at least on the in the know part one today just to address uh the the three topics that we think have really unnerved the markets: AI hype, crypto meltdown, and macro macro statistics, but more macro um variables like Fedspeak. Uh, so let's go through each one of those.
Uh, and we'll start with the uh AI hype. So you can see from this first chart here uh what our predictions in December of 2024 were for really the tech stack uh as AI uh was evolving. Uh, and you can see here that we saw rapid growth in all three areas of the tech stack: so the infrastructure layer, the platform layer, and the application layer. Uh, that was true uh in the four years ended or five years ended '24 and our expectations uh for the next five to six years. Now, what we also pointed out below is that there were share shifts taking place place in terms of incremental growth among these three. Uh, and you can see that the big winner was platform as a service, think Palunteer, one of uh the most important companies in this AI revolution. Uh, and its growth rate, US commercial alone uh last quarter was 142% in revenue growth. Uh, no other software company is coming close to that, at least in the public equity markets. In the in the private markets, of course, uh you've got the frontier model players um running away with 250 to 850% growth rates. Uh, think OpenAI and and Anthropic in the last year, year and a half annualized. Um, you can see the big loser we thought was going to be the application layer, uh largely software as a service. Now, if there's something wrong with this chart, uh you can see it is in our projection for software as a service going forward on the top there, uh 20% plus. Uh, it's not going to be anything like that. In fact, uh we do believe that software as a service is becoming a victim of this AI revolution. There will be uh consolidators uh and and software as a service providers who do survive this route. They will consolidate the market. Uh, but we, we underestimated ourselves. We were saying they that SAS was going to lose share. We didn't realize how much, how soon, and that G gives you a sense of uh how quickly uh this revolution is happening. Uh, 142% for Palunteer uh uh US commercial in one quarter. Uh, we didn't expect that either. In fact, our kagger, our compound annual rate of uh revenue growth assumption, which we thought was enormous at the time for the next five years for Palunteer itself, uh is 55%. That could be wrong too. Maybe it's too low. Uh, so here we are.
Okay. Next, uh uh what about all of these capital spending announcements uh from the hyperscalers, especially? This is unnerving. Uh, the the the mag six uh or investors holding the mag six. These investors, many of them benchmark sensitive. What does that mean for people who who don't understand what that means? Um, the market since the tech and telecom bust and uh so that was the early 2000s and the uh '08-'09 meltdown has become uh much more quantitatively driven, which means uh that many portfolios are guided heavily by the indexes out there. Now, to give you a sense of that, um Amazon and Nvidia in uh the NASDAQ 100 uh are in the 3 to 4% range uh in terms of the holdings. And uh what benchmark sensitivity means is uh portfolio managers will say, okay, uh I am betting against, let's just say Nvidia for argument's sake. Nvidia, I think is at a 3.4% position in uh the NASDAQ 100. And so those portfolio managers will take their position down from maybe they were optimistic on Nvidia and they were at 3.8. Well, they'll take it down to three. And to them, that's betting against uh Nvidia. Uh, and uh and to them, that's also taking a big risk because they're being compared against other managers who are doing much the same thing. Uh, as you know, Arc Invest doesn't think that way at all. Uh, we are uh focused on the future and uh we're very focused on pure plays, not that I mean, we do hold both Nvidia and um and Meta uh in our funds, but for the most part, certainly in our our flagship strategy, uh we're well below that weight, well below in the one to one and a half percent range, just to give you an example of uh how how portfolio management has evolved over the years in the traditional asset management world.
Um, so I think that the shareholder base holding the mag six, they have gotten so used to uh cash hordes and massive key uh free cash flow generation, that this is a big shock to them seeing these hyperscalers uh take uh their capital spending budgets up 100% in one year. Uh, I think Amazon announced it would go from roughly 90 billion uh to uh 180 billion next year or this year, 2026. That's shocking to people. And uh we're seeing some of the mag six uh uh raising debt, even that has been anathema to this group. Uh, so we do think we we're seeing a a shareholder turnover. Uh, not massive because of what I just described about benchmarks, um, but uh we don't think necessarily that they are doing the wrong things, meaning the hyperscalers. Uh, we think Google is doing the right thing, um, and all of the Amazon, all of this spending, we think is the right thing to do. But this is not what traditional portfolio managers who are benchmark sensitive uh are used to grappling with. Um, we grapple with it uh all the time.
But here you can see what we, this we put in um our Big Ideas 2026. Uh, you can see and what we're trying to illustrate here is uh investment in data center systems. Now, I know you've probably seen the $650 billion uh spending, capital spending boom that newspapers are highlighting. Um, that includes much more than data center systems. So, just want to make that clear. So you can see here that uh data center system spending growth was really mid-single digits uh at an annualized rate uh for many years, and it and it kind of stayed in the 100 to 200 billion range until the ChatGPT moment. And since then, you've seen a compound annual rate of growth in at 29% % uh uh I think with the announcements today, 100% in one year. Uh, we're moving into overdrive. Now, you can see from our forecast uh that we've been expecting a 30% compound annual rate of uh increases since '23. And um, I think if we were to push this forward to '25 now and do uh a five-year uh kagger, it would be in the 25% range. Uh, so these recent announcements uh certainly um are proving this forecast uh right. And then you can see, just to throw in uh another perspective here, we also see that the chip market is is going to um transform or evolve, I should say, in the next five years, uh a little bit uh from e GPUs taking share uh to ASICs taking share. And so, uh, we've, um, highlighted that there could be a shift from the likes of Nvidia's chips, uh, AMD's chips, uh, to Broadcom's chips. Now, uh, we think that there's going to be a massive appetite for all chips. But did want to highlight this share shift as well, uh much like we highlighted the shift to away from SAS uh toward platform as a service.
Now, this spending, is it justified? This chart, Brett Winton, our chief futurist, um prepared for Big Ideas 2026. It highlights the technology revolution investment cycles here. And you can see the railroad in the 1800s. Railroad spending as capital spending as a percent of GDP got into that five to six percent range. Then you go to uh the automobile in red there. Uh, but we have multiple flat platforms evolving at the same time, and this is additive. So maybe with telephone, electricity, and the automobile, uh we would see capital spending um up at that cumulatively up at that uh five to six percent range. And then you get to the internet, um, the tech and telecom boom, uh, and you can see it looks like it pales in comparison to those other two. But we do have an additive, uh, situation going on here. And you can see that adding all of those together is maybe the equivalent of just the automobile cycle uh in the early 1900s. And then here we are, and we, we believe these are our forecasts, and they have been our forecasts, and we haven't changed our forecasts in the last year. But we've got something additive going on here as well, um, that could accumulate to 12% or more of GDP. Now, just to give you a sense of uh or some perspective, I remember in the bubble, te te telecom equipment spending on the part of some uh providers got up to 25% of GDP, and as I always say, too much, too soon. And uh and we ended up with a lot of dark fiber, it didn't go live for a very long time. One thing we can say about this cycle is the demand for GPUs is voracious, and they are landing in these data centers, which are saying, we need more capacity, we need more chips. Those GPUs are not dark. And I'm stealing that from uh Brad Gersonner at Altimeter. I thought that was a very good analogy. The fiber went dark. GPUs are alive, and there's a big scramble for them.
Now, the risk here is, and you know, I have an economics background, and I'm always looking for those risks. Whenever I hear the word shortage, uh, I usually, um, believe that there's going to be a glut, an inventory glut at some point. And I'm sure there will be in the chip space. But I don't think it's going to be for a long time. Uh, you need to listen to Elon Musk, who's talking about uh chips and power as uh, you know, for as far as the eye can see u being in short supply. And uh, now that he's thinking about putting data centers into space, perhaps even more so.
So on this next chart, this this is getting uh to investors. Um, so you can see on the left here, the the tech sector capital spending as a percent of GDP, and you can see that it has gotten up to uh be bubble days. And that's a little bit unnerving to investors. And here's why. Many of the most seasoned investors today um were very young when the tech and telecom bubble took off, and now they are the gray hairs. And I'm friends with a lot of them. And many of them are saying, been there, done that. I'm leading uh this organization now, and I'm going to protect it from this bubble. And we understand that. But as I just described, we have really just started this journey, and we expect the capital spending cycle to go much further for much longer. And uh the reason is the technologies are ready. You know, back in the '90s, we didn't get the cloud even until 2006. But there were visions of, you know, eyeballs uh uh in 10 years being the source of valuation today. We didn't get the cloud till, or or back then, we didn't get the cloud until '06. We didn't get the first big breakthrough in AI until deep learning, ImageNet in 2012. Even bigger and more profound, transformer architecture in 2017. And even off of that, we didn't get ChatGPT, the aha moment, until the end of '22, early '23. So we have just begun this journey.
Now, one other difference, very important, you can see from this chart, is the valuations today are much different than they were at the peak of the bubble. So the dash line there, you can see are uh that is the valuation of the mag six. And you can see how crazy the valuations did get at a time when the technologies weren't ready and the costs were too high. So the multiples being paid today are nothing like back then. And in fact, since this chart was done, the multiples have dropped even lower because it is the mag six, and and fears around what are they doing with all of this capital spending, return on invested capital is going to go down. That's why I own this stock. And that's why a lot of quants or quantitative research is a little bit um, I would say, a little dangerous here, because a lot of it's momentum-driven. And here we are. There's a great disruption taking place called AI, and these companies must invest in it if they want to remain relevant, and they must seize the moment. But short-term oriented shareholders don't really want to hear that.
So on this next chart, we're heading into crypto. And we'll loop back to AI at the very end. Crypto here, I think the disconnect, and this is also from Big Ideas, is, hey, look at what the gold price has done in the last few years, and look at what's happened to Bitcoin. Now, this is a log chart. So, um, you can see, uh, gold has had a massive move to the upside. This chart ends in at the end of '25. And of course, uh, I think last night we got down to roughly $60,000 in in Bitcoin. So, really cut in half since its peak in October. What is going on here? Isn't Bitcoin supposed to be digital gold? If it's digital gold, why isn't it behaving like gold? And a couple of points to make here. If you do the correlation between Bitcoin and gold, their returns, uh, since 2019, and and I put this in, uh, my New Year's letter, you can see the matrix there if you'd like. The correlation has been 0.14. One is perfect correlation. And here we are at 0.14, very low, almost no correlation. So that is very important to understand.
Um, what you can also see here, though, is gold tends to precede a big move in bit uh in Bitcoin. You can see the circles there, and we'd have a gigantic circle around what has happened recently. Um, and and so maybe that will happen again. We actually think it will. I cannot tell you how many bare stories I've heard about Bitcoin uh in in the last few weeks, last few months. And just go on to X, and you'll find all kinds of reasons why you must sell Bitcoin. And these are the the periods that are kind of interesting. We post our trading at the end of every day, and you'll say, you'll see that we had been nibbling at our our crypto-exposed stocks. We didn't know how far this was going to go, and maybe it's not over. Not promising that, but um, uh, there was so much negative sentiment that I, I felt, okay, you know, we need to leg in and lean into this if our research is correct, that this, um, this Bitcoin is three revolutions in one. It's a global digital private, very important, rules-based, even more important, monetary system. And that's a very big idea.
The second revolution is it is a technology. It's a new technology. It's uh it's really adding another layer, I would say, to the internet so that we have um native currencies native to the internet that will be able to handle uh agentic commerce and uh other other such agentic movements taking place out there. That's a very big idea. And then, and it won't be, of course, just Bitcoin, it'll be Ether, importantly, Ether, Solana, maybe Hyperliquid, we shall see. But what's interesting about this period is the big three have been identified, and those are Bitcoin, Ether, and Solana. You know, the worry historically has, how many of these things are there going to be? Is this going to be a very fragmented situation? And we're learning, we, we're learning who the big winners are are going to be, who the big winners are. And I think that's a very important.
And then the third revolution is this is a new asset class. Bitcoin is leading the way. It is the mo most secure of all the crypto uh currencies out there, backed by the largest computer network in the world. And so really an anchor to um an asset allocator looking to move into this new world. And why would an asset allocator do that? Well, if we're right, and well, the it's a fact that at least over the last five years, the correlations have been, the correlation has been as low as 0.14. The key to successful diversification um for asset allocators is um adding new assets with very low correlation to all of the other assets in the ecosystem. And uh this one qualifies. Why do they want that? It is because with that low correlation, introducing it, um risk-adjusted returns go up over time. And so I do think institutions are serious about this new asset class. They may have been reticent uh to move in because they were hearing about the four-year cycle, the four-year cycle. And lo and behold, and I'll, I'll hats off to Chris Berniski, who has been uh very right in in calling for whether it's the four-year cycle or whatever, just this downdraft. And so it's close to, it's closing in on the low point that many uh technical analysts are suggesting as the place where it should stop. These, these are very volatile moments at at or near the bottom. But uh we usually do look back and say, wow, I wish I had caught that low. This V-shaped recovery um uh has been quite significant. Not promising. I want to make, I want to make sure that there are no promises here. But it does seem like the stars are aligning. And um, I guess I'll move on to the next one.
So what has been the biggest bugaboo in as it relates to Bitcoin or the worry that we've heard about more than anything else in the last few weeks? Because it's all about uh quantum computing. The last few weeks have been uh, you know, lots of stories about quantum computing. And I would say, uh, the the large shift in Bitcoin, the OGs selling uh or just shifting uh wallets to more quantum-resistant wallets. So I think that's really uh caused a lot of concern. And here's Michael Sailor's uh answer. Or I mean, he was very reasonable last night in his discussion around quantum. He doesn't think it's going to be an issue for 10 years or more. But he said, here's what we're doing uh in terms of our commitment to Bitcoin security. And and so you can see some of those um those moves now.
But I also wanted to show our research. Again, this is out of Big Ideas. And this is when the quantum controversy uh had quieted down last year as it relates to Bitcoin. So Brett Winton, uh in as chief futurist, has um has a a big focus on quantum because we can't miss the next big thing. It's all we do is disruptive innovation. And here I thought was a compelling analysis uh based on what has happened with quantum computing so far. And you and Google really leading the charge, that little black line on that chart. So you can see that um it has been able to double the number of qubits over four years, which is slower than Moore's Law in the semiconductor space. And of course, there is an error rate consideration lowering that as well over four years. So if if Google does not pick up speed, the red line tells you how long it will take Google to and uh quantum computing to really threaten Bitcoin, the Bitcoin blockchain as we know it now. So we go into the 60s there. Now, if you make the assumption that Google is going to pick up the pace, and that's an interesting assumption, because typically we learn a lot in the early years of a new technology. You know, how how quickly are the units increasing and driving costs down? That's the learning curve associated uh with new technologies. But let's assume it, assume it does pick up the pace, doubles the pace to Moore's Law. Okay, then uh we have a problem in the mid-40s. So I don't want to make light of it. I do think it has caused movement uh uh among of Bitcoin uh between wallets, between and among wallets. Are they leaving the crypto ecosystem? Some say yes. Maybe they're diversifying into other crypto assets as um as the DeFi ecosystem evolves here. I don't know how likely that is because these are Bitcoin OGs. They've been there uh forever, and so we'll see. Maybe they're diversifying because they were HODLers and now they're 10, 15 years older and they have families, and you know, their risk profiles have gone down. They should they should diversify. Any financial advisor would say so. We'll find out. But um, uh, I thought Brett added some good perspective here.
Now, going into economics, and again, I'll, I'll whip through this last part because I'll do a a more in-depth in the note next week when the employment numbers do come out. So here you can see the liquidity index. This is Dan Rodriguez's chart, and you'll can you'll see liquidity dropped to uh dropped to a low level in November. Why was that? It was the government shutdown, and we thought, okay, this is going to be short-term, liquidity will pick back up, and the market uh will will take off again. And that is what happened. But recently, you can see that it is back down. And what's happening there? Well, we had the threat of the shutdown, and now we actually have the shutdown, partial shutdown. And in mid-February, I think it's, I think it might be February 14th, there's going to be a vote around uh the Department of Homeland Security's budgeting, and you know, it may shut down more completely then. So, I do think there is somewhat of a liquidity drain. And then the other thing that's happening is the the Fed was supposed to stop quantitative tightening in early December, and it, and it did for the most part. But its provision of liquidity to the system is lower than many analysts expected. In recent days, in recent days, it has started to pick up. You're seeing some of that now. And I do think the Fed and the Treasury are beginning to respond to what has happened to the financial markets. This administration uh does think, does believe that the financial markets provide important signals, and so they are probably assessing the liquidity liquidity uh conditions out there and um, and moving accordingly. So I do think this one is probably moving in the other direction.
Here you go. You've seen this chart many times. Metals to gold. Uh, uh, we've learned something very important, and I'll spend more time on this in uh next week. But what surprised me, and this is up to date, up to date, is with the fall off in gold recently, that the metals to gold ratio has not picked up. That is really surprising to us. We've learned in recent weeks uh why the treasury yield uh has been levitated. This very tight correlation between the metals to gold ratio and the Treasury yield broke down starting with the Fed tightening, and you know, it's still gaping wide. We're learning and believing more and more that this is the yen carry trade unwind. What are they doing? Selling Treasuries and buying Japanese bonds. Why? Because Japanese interest rates have gone up, and so the the US Treasury bill is less of a, less interesting. It's still above the short-term rates in Japan, but less so, much less so actually, and could continue in that direction if the current prime minister, I, is, uh, she just called a snap election, and I think it's for this weekend. And if she gets a large majority, larger majority, she will amp fiscal spending even higher, and that will probably take their interest rates higher. Nonetheless, the purple side is puzzling. I think it has to do with deflation uh coming out of China uh and impacting the global ecosystem. Sure, we've seen some metals go up, but this ratio going down uh even further as gold goes down is most surprising here.
I just want to go through, I, I presented this chart at the last in the no. So we'll, we'll just give you a sense of how crazy it got a couple of weeks ago, gold going parabolic as a percent of M2. And many people pushed back on our use of US M2. You can't get global M2 back into the '20s, and the US still is the reserve currency. I don't think the message would be much different with a global M2. So you can see it hit an all-time high, higher than the Great Depression, at uh 1.66 versus 1.76 versus uh 1.71 in the Great, Great Depression in the '20s. Today it's at 1.54. So we do think the bloom is off the rose, but it's still at very extended levels. And we wouldn't be surprised if gold continued to come down uh to Bitcoin's benefit. We shall see. And also, I think uh Bitcoin has been hurt uh by all of the funds flowing into AI that became the interesting uh investment opportunity, and that one has become a little shaky. So I think, you know, these three, gold, AI, and Bitcoin are interrelated in that way.
Now, gold is a hedge against inflation. Look what Trueflation is telling us. It's telling us inflation is resolving to the downside. It is not moving up. And there are a few things going on. Gasoline prices until very recently, maybe housing prices, and that's going to continue as rents come down. And even potato chips, Pepsi, Frito, Doritos cutting prices 15%. And Hims & Hers, uh, although the FTC, or no, it's yes, I think the FTC is going after them, cutting the price of GLP-1s to 40, I think at $49, whereas I think Novo Nordisk has cut its entry price or introductory price to $149, and then it's going to move them up to $199. These, this is deflation. This is deflation. And um, we, we're going to see a lot more of it. We believe.
Unemployment rate. And of course, all of this is uh is trying to figure out, okay, what's the Fed going to say next? I was uh very surprised that uh Federal Reserve Governor Bostik said employment is very strong, and he's looking at the overall unemployment rate. And when you have 1.3 million baby boomers leaving per year, you know, the overall unemployment rate is probably going to stay fairly low. And we've also got um, uh immigrants who are uh exiting, and we don't have as much of an inflow of immigrants. So the unemployment rate is going to stay low overall. But you can see what's happening to the 16 to 24 year olds. Their unemployment rate has gone up from roughly 7% to more than 10%. The last time I presented this, it was 12.2%. I think there were downward revisions, and we'll, we'll get to the bottom of that, but still 10%, and the duration of employment almost a half a year. So I do think that the the concern about employment will build as more and more Fed governors understand entry-level jobs are disappearing. And so a very important demographic is not getting jobs. We do think that's going to change, as a lot of these young people go out there and go for it in terms of starting their own businesses, because the barrier to innovation, the barrier to uh starting new businesses is dropping precipitously with AI.
I'm just going to end on two charts. Uh, uh, once again, we're very optimistic about GDP growth. We think we're going into a super, a productivity-driven boom. And just reminder, the step function change in GDP growth when we had the last big technology revolution, which was the early 1900s. That growth rate stepped up from 0.6 to 3%, five-fold. And we're saying 7 to 8% by the end of the decade. So I think that's conservative given all of the, well, given the technology revolution underway.
And finally, this is how we think uh that that innovation is going to play out in uh the equity markets, the market cap. On on the right-hand side there, you can see what happened to uh railroad stocks as a percent of um as a percent of the total global equity market cap, 65%. You can see where we are now, and we think this revolution is is even bigger, as I mentioned, and really just starting in its first few years, at least. And you, you can see where we think innovation is going to go as a percent of global market cap. Again, we've just begun. We don't think AI hype is where we are at all. We do not think we're in the equivalent of the tech and telecom bubble. We're very happy that many people are worried about it. In the tech and telecom bubble, nobody was worried about anything. In fact, I'll remember clearly, Jeff Bezos at Amazon, and I think this was in 1999, he with some glee announced on on Amazon's earnings call that Amazon was going to lose so much more money in the next year because Amazon, Jeff had decided this opportunity was so much bigger than anyone understood, and by the way, he was right. It was going to take much longer that he had to up the capital spending budget. And what happened that day? The stock just soared. Soared. And one of um our panelists on our brainstorm this week reminded me that um that a a journalist asked Jeff, "Well, what is, how do you spell profit in this new world? What does that mean?" And he spelled it P R O P H T. So that's how cavalier uh CEOs of innovation-based companies were back then, those driving the innovation. And then, of course, they paid a heavy price for that over the next three to four years. The fear in the market today, when I see Microsoft going down 12% because it's doubling its capital spending over the next year, or Amazon going down 8, 10% because it's doubling its capital spending over the next, completely different response, completely different um investor mindset. I think there's real fear out there. That was one of the reasons gold went up, fear of everything. And I just think, and this is why I wanted to do the call today, I just think we're in a completely different era. No resemblance to the late '90s. I was there. And uh, I think we're in for very exciting times. And don't underestimate, uh, the growth rates that these new technologies are going to spur. Uh, we think they're going to be super exponential growth rates. And Palantir is a very good example of that. No one expects the 142% uh US commercial revenue growth to be sustained. We don't either. But you know, if it were able to sustain, you know, 70, 80%, which is higher than what we have in our model, um, that stock would be a screaming buy today. So we are assessing all of this as everyone else is in the financial world. It's a really exciting time. It's a confusing time. There's going to be a lot of chaos. You saw a lot of auto manufacturers basically shutting down their EV efforts. They couldn't scale them uh quickly enough to capture the cost declines. They're giving up. And one of the most interesting uh ideas I heard on our brainstorm today was the idea that, well, who's going to roll the auto sector up? Normally you think it'd be an auto company, one of the big ones. Someone suggested no, maybe it's Google. They need cars.
So with that, I know we have a, we have a relief rally today in Friday. And no one knows what the future really holds. But um, I just wanted to be in front of everybody because I know how harrowing it must have felt over the last week. And just wanted to reinforce our point of view, uh, because we think a lot of the data that has come out recently does just that. So, I'll see you next week.