228: Best Of: Technology and AI Investing Trends - A Stock Picker's Take

Welcome to The Bid, where we break down what's happening in the markets and explore the forces changing the economy and finance. I'm Oscar Pulido. And I'm Stevie Manns, The Bid's producer. We're continuing our summer series with a look back on some of our favorite episodes. Last time I picked a favorite, and this time, Stevie, it's your choice. So what do you have for us today? I have chosen our episode with Tony Kim. It was called Tech and AI Investing Trends, A Stock Picker's Take. I do remember that. I remember there was a visit through Silicon Valley that we chronicled during that episode. So tell me a little bit more about why you chose this as one of your favorites. So as you know, I'm a self-confessed AI nerd. I'm just so fascinated by it. You might be aware that we use it on The Bid here and there. Sometimes it helps me read through some white papers. It might help put some questions together. We sometimes use it to translate episodes of The Bid. And we might use it on some of the backend metadata stuff for the podcast as well. But also I love in this episode, Tony Kim's passion for technology and history really came through. We were at dinner one evening after the tour. And Tony was telling this story about the ancient Egyptians and the engineering. And he was so animated by it. I thought this is going to be a great episode. We hadn't recorded yet, but I knew that it was going to be great. And I loved how he broke down the AI tech investing stack from chips to applications. And why this moment in AI is just so transformative as a period in history. You definitely got the sense when you were listening to Tony that this is somebody who knows this space and has been following it for a while. And also hearing from some of the executives on that episode, the chief executives or the leaders of some of the companies that were visited was good context as well to hear from companies that maybe in some cases you don't know who these companies are. You haven't heard of them before, but it was a good opportunity to start hearing about who are some of these companies in the ecosystem. With having said all of that, why don't we go ahead and take a listen to the episode? Tony, thank you so much for joining us on The Bit. Thanks for having me. It's a pleasure to be here. So earlier this summer, you hosted your 11th annual tech tour where you and 36 other Black Rock colleagues took a bus more than 300 miles over five days to meet with leaders of over nearly 30 technology companies in San Francisco and Silicon Valley. I have to say that sounds like a pretty fun trip. Not surprisingly, AI was a topic you talked about a lot. This is a space and a topic that has been changing quickly over the past year though. So I'm curious, how has the conversation evolved over that past year? Yeah, as you mentioned, this is the 11th year we have done this. The first nine years there was some talk of AI, but generally very little, very company specific. Last year, 2023 and this year, AI has dominated the conversation. In fact, it was a topic of discussion in every single meeting. What also is emblematic of especially this year is, as you mentioned, the rapid change of AI and how AI has taken over the entirety of Silicon Valley. There's been a huge pivot and this has dominated the strategies of every single company that we visited on this tour and basically across the entirety of tech. What's changed is in the last 12 months is that intensity has come even clearer into focus. The strategies are becoming more apparent, more definitive. The commitment and focus is higher than it was even last year, which was like really the first year we really saw AI come into its own. It's fascinating to me that in the first nine years of your tech tour, AI was barely a mention or you said it wasn't a focus, but now the last two years, it's really the only topic of discussion. Yeah, it is really the only topic of discussion because in the prior nine years, AI was more a tool, something in the background. Operationally, a few companies were using it. It is now at the forefront of every company. It's really changing the foundations of the software industry, semiconductor industry. It's cascading to the power and energy industry. It's also changing the very product and services that all tech companies are building to. That's the sea change. And it's been remarkable how that's changed in 24 months. So Tony, it turns out we were able to join you on this tour. And as you were meeting with different industry leaders and company leaders, we actually were able to get some commentary from them. So first, we'll hear from Arsalan Tabakoli-Shiraji, co-founder of Databricks, who has been an AI believer since his company was founded 10 years ago. He also is struck by the recent pace of change. The pace of change has been incredible. Let's say a year ago, 18 months ago, nobody really was even talking about AI. And now all we talk about is, what are GPUs going to be used for? What are all the different systems? How do I harness AI? It's been a while since I've seen something that has so much excitement from industry, yet still is so nascent that it's in the research phase. So Tony, clearly in public markets, AI has been powering the phenomenal run in stock markets, particularly companies like NVIDIA and a handful of other mega cap stocks. But the investment opportunities certainly transcend that small group of names. So you're a tech investor. Where are you finding the most compelling investment opportunities? On one hand, one dimension we should think about AI is there's a stack of products and services. At the bottom of the stack, there is the chips and the infrastructure, the cloud infrastructure. A lot of the investments, especially last year and this year, have been to build that foundation, basically a rebuilding of the internet, if you will, of a new computing infrastructure, which is AI. There is then another layer, let's call it the intelligence layer, which is the data and the models themselves, the foundation models. Those are being built and those are increasing in capability. And then on top of that, there are software infrastructure and software applications and then services and solutions that combine all of this with the AI intelligence. That is now also starting to be put together of new applications that are leveraging AI. So you have these three layers, this infrastructure layer, this intelligence layer, and then this application layer. The investments, you start from the bottom and you move up to the top. And in 2023 and 2024, a lot of the investment and a lot of the stock price reaction has been at this bottom layer of the stack. This is an AI tech-led new bull market is what we like to say. Jensen Huang touted the extraordinary demand for AI chips on Wednesday. Demand is so great that delivery of our components and our technology and our infrastructure and software is really emotional for people. However, it takes time for us to move up the stack. And as we progress into year three, year four, year five of this thing, we will continue to start seeing opportunities of companies in that intelligence layer. And of course, finally, in the application layer, which has not been as a beneficiary of this first wave of AI, which we saw at the bottom layer of the stack and infrastructure. We as investors, we are looking across all of these layers of the stack. The market will evolve over time, but we are seeing at least in 2023 and 2024 for sure, a huge focus and emphasis of building this foundation of the AI infrastructure. Tony, you talked about the tech stack that as you see AI, the different components of the AI ecosystem and, but maybe tell us a little bit more about the end markets and the end users that benefit from that AI spend. I say the first end market is around these big cloud AI computing companies building this infrastructure and all those companies that are associated with that, that sets the foundation, that infrastructure spend then gets leveraged into, let's say the first end market will be consumer. And it just so happens that many of these big infrastructure companies have big consumer businesses, search, social e-commerce shopping. And I think you will start to see AIs come into the consumer market. And this is a market of what I would call personal assistance. The second big end market will be around businesses, enterprises that will then again, leverage this infrastructure spending and bring AI into companies. The third is the end market for the real world. Cars, airplanes, the military, robots, et cetera, the instantiation of AI into the physical world. These three additional end markets, again, I would say we're in very early stages of adoption, first inning or less in many cases. But they all leverage the same infrastructure that is being invested right now at these big foundational models and foundational compute layer. And then that gets amortized and leveraged across all of these ultimate end markets for consumers, businesses, and the real world. Tony, when it comes to CapEx spending or capital expenditures, essentially what corporations are spending on that new computing infrastructure. I think that's the term that you used. Google has said that the risk of under-investing is far greater than the risk of over-investing. A lot of the big tech companies have committed to massive amounts of capital expenditures towards AI. Yet, it may still not be enough as Hock Tan, president and CEO of Broadcom shared. The amount of money that we involve, we haven't even began to quantify. And one could imagine the level of spending required both on computing engines, on software models, on infrastructure and power in particular, could possibly be larger than what we all are thinking of at this point. And that's why I say we're probably underestimating the amount of dollars in CapEx, an investment we would need to make in order to reach that aspiration goal of what we call AGI, convergence in artificial intelligence. Tony, are you worried at all that the capital expenditures on AI might not deliver the return on investment that people think it will? In a singular answer, no. However, Wall Street and investors are obviously asking that question now about ROI of AI. The issue is the two and 10-year question. Maybe in the next two years, we are overestimating the impact of AI, but in the 10 years, we might be underestimating the impact of AI. I would generally agree with that. If we are trying to match investment dollars in and profit out, in the near term, that is a mismatch. Clearly, companies are spending more than the dollars that they're getting out right now, the two-year problem. But in the 10-year duration, the potential benefits and returns that we can get, or maybe we're underestimating that. So I would generally say I would agree with that statement. You can't match this spending with the revenue and returns linearly, one for one. But as an answer to your question, I'm not really concerned about that. Your comment about Google and the risk of underinvesting, I would also agree with that. The big three hyperscale, Amazon AWS, Microsoft Azure, and Google, all three have very, very large cloud computing businesses that they've built over the last 15 years. That has been the engine of growth for many of those companies. But those cloud businesses represent the cloud of the pre-AI era, the SaaS revolution, the cloud computing era of moving compute from customers' premises on-prem to the cloud. This has been a tremendous tailwind that's driven these companies for the last decade or more. AI has come, and we've got to rebuild a completely new computing infrastructure. AI then will be instantiated in all these new applications and markets. So if you are one of these big three companies, and if you do not invest, and you so happen to pause or slow down, and your competitors continue, you run the risk of falling behind. And if one of your competitors also builds an AI capability, a technology breakthrough that leapfrogs you because you're not investing, then these are existential potential risks to the current business as currently constructed in the prior era. And so in a way, it is not only the potential long-term benefits from AI in terms of the returns that come with that. In my opinion, there are certain existential questions. If you do not continue to invest, you run the risk of falling behind, which then impacts your current core business. So that, again, goes to this two-in-ten year. Are you only trying to optimize for the near-term profit, or are you ensuring your position in the long-term future? Just one more thing about the CapEx. It is a lot of money. I did an exercise. I was adding up the top 10 spenders just this year. It's roughly $270 billion of CapEx, of which the top four is $230 billion of CapEx, plus or minus. So there is a lot of spending going into this year. But when you look at the balance sheets and cash flow and profitability of these companies, they can afford it. It doesn't seem excessive to me. The other thing about the spending is it's a lot of money, but it's also the moat. This capital that is required to build these most advanced AIs for these companies is also the moat. And so this very nature of CapEx intensity is also a feature. It's also the moat, the defensibility. That all goes in hand with this long-term strategic and existential risk. But it creates, to me, bigger, longer, enduring moats and duration. We heard similar excitement from Rodrigo Liang, co-founder and CEO of SambaNova. You look at all these innovations that are happening across the board, whether that's accuracy, whether that's multimodality, whether that's performance and speed, whether that's efficiency, you see this innovation across all of these different ways that the models are operating. And I think you're going to only see it accelerate. It's a tremendous time for innovation. It's a tremendous time for technologists. And we're really excited to be in the middle of it. And Tony, just hearing you talk, you obviously have a very long-term lens on the benefits of AI. You talked about that 10-year window. It makes sense now why in a recent article in The Atlantic, you were dubbed an AI optimist. It's clear just from the comments that you're making that I think that's a fair label to attach to you. What would you say then to the AI skeptics or those that are questioning its viability? In terms of the AI skeptics, I think this goes back to this two and 10-year discussion. From what I see, many of the skeptics are wanting this perfectly matched investment with return right now, today. You're not going to get it. It's just not matched. Then that would shift the discussion to then do you believe in the longer-term returns and benefits? And I think I would say one of those is the strategic and existential because you can't just assume that the current status quo will be sustained. The second is then what are the long-term benefits that could come and the returns that could come with AI? And here, it does require some leap of faith, some imagination, some creativity of thinking of what are all the possible outcomes and changes that AI can have to business and to society and to productivity. I do believe it is the means to an end. The investment justifies the potential return and these are difficult to counteract the skeptics right now because we just don't know when and what those will be and what magnitude. There are, of course, a number of companies who are innovating by supporting the AI revolution, not by developing large language models, but instead things like quantum computing to assist with faster AI computations. And we heard from Fariba Dinesh. She's COO of SciQuantum, a company building the world's first useful quantum computer and asked her about the potential for quantum computing. I think the most opportunity for quantum computing to contribute to humanity is climate and drug development because chemistry is quite complex and chemical problems are not something we can simulate today and computing has contributed to that very little over the years. So with quantum computing, there's enough compute power where you can actually simulate these very complex mechanisms that happen with bond energies, etc., etc. Quantum computing is uniquely good for quantum type mechanisms, which, of course, chemistry is all about that. So, Tony, speaking of quantum computing, which I know is a theme that you've been an early supporter of, what are the investment opportunities in this space? It feels like it's something that's been getting less attention than straight AI-focused companies. Absolutely. Quantum computing represents a next generation of computing. The last 50 years we've been and we currently are still in the classical computing era. And I would say these next five years is going to be absolutely incredible in classical computing driven by AI and what NVIDIA is spearheading in building these clusters of AI computing power. And this is all classical computing. And we are barreling toward AGI and these super nodes with this unimaginable computing power, but they're all classical, what we call binary classical computer. However, parallel to this, it's been a long while coming and we are now at the precipice of the dawn of, let's call it, a parallel. It's not replacing classical computing, but it sits adjacent to it, a quantum computer which functions very differently to a classical computer. And I always say it's more mimicking nature. You look at the real world and nature and it's driven by quantum mechanics. And to harness that capability, we can build computers of unimaginable computational capability to run very specific kinds of problems that classical computers cannot solve. Problems such as advanced simulations, optimizations, security, encryption, factoring numbers, these problems classical computers cannot solve. And we are within the next five years at the precipice of breaking through a utility scale quantum computing. So if you just fast forward, let's say to 2030, on one hand, on the classical computing side, we will have probably computers and AI supercomputer nodes that could approach a million GPUs per node, creating and enabling the building of the most advanced AI models humanity's ever seen. Parallel to that, we could have a breakthrough of utility scale, super quantum computing and through a company like SciQuantum that would sit parallel to this AI supercomputer, classical computer, and solve unimaginable problems and also create and help generate unimaginable data to also help train the AI supercomputer. So what is coming, I think, in the next five years, let's call it the end of the century, we could have exponential scale up in terms of computing power, both classical AI as well as quantum. And that is what is exciting. And if you could have one company that breaks through, you could have an open AI-like moment for the quantum computing industry. And Tony, perhaps one final question, as you're talking about the future of what AI could bring, it reminds me that you're not only a prominent technology investor, but that you also have a passion for history. So tell us a little bit about why these two passions intersect and relate to each other. I love history. It might not seem obvious. I have an engineering background, but my true passion has been history. I also look at historical figures, either via Napoleon or Caesar or Alexander the Great and their strategies. Also, I look at the history of math, physics, and computer science, the quantum mechanics, the history of encryption and the computer with the Enigma machine, trying to crack the Enigma machine. All of these kinds of past historical events, historical breakthroughs, they're very formative in how I also look at the future and the technology. And actually, I quite live in analog world, even though I am living and working in the most advanced technology, I'm more of an analog person. Well, something tells me somebody is going to be reading about you in a history book one day, Tony. You were at the forefront of the AI revolution and helping people see what was coming. Thank you for sharing your insights and thank you for doing it here on The Bid. Thank you very much. Well, it's great to listen to that episode again. The insights on CapEx and the AI stack that you mentioned, Stevie, were so sharp. I love the historical lens that Tony brought to the conversation. He mentioned Napoleon, Caesar, quantum computing, which is sort of like coming soon. You're hearing a lot more about it now. And Tony was telling us about it last year. Really, only Tony could tie all of that together. So Stevie, it turns out you're heading to San Francisco again very soon to do this again to get some clips for a new episode in the fall. Is that right? That is right. Spoiler alert. Yes, I'm going back to San Francisco. And last episode, we were talking about Davos in the Chile Alps. So it's a nice change to go somewhere sunny. I'd never been to San Francisco until last year, and it was so fun. And going to see these companies was just so eye-opening to be on the ground in Silicon Valley and like this backstage pass that was really unique. We got to see some very cool demos, some robots, and I'm just very curious to see what trends Tony will pull out for us in this next episode. Well, I'm excited to record that one as well. When you're back from your trip, I look forward to hearing about not only the weather, but also what some of those conversations with those companies were like. And I will miss you, but I'll bring back some great clips for you. But next up, Oscar, it's back to you for our final choice for a Best Of episode. So do join us again for another look at one of our favorites. And again, we hope you enjoy revisiting these episodes as much as we do. We'll be back soon with fresh conversations. But until then, thanks for listening to The Bid, and subscribe to The Bid wherever you get your podcasts. I'm Oscar Pulido. And I'm Stevie Manns. See you next time. Reference to the names of each company mentioned in this communication is mainly for explaining the investment strategy and should not be construed as investment advice or investment recommendation of those companies. For full disclosures, go to blackrock.com forward slash corporate slash compliance slash bid dash disclosures.