230. AI At The Frontier – A Stock Picker’s Take on Tech and AI investing
Episode Description:
The world of artificial intelligence continues to profoundly impact the stock markets and create investment opportunities. Despite a brief setback earlier this year, AI continues to push the boundaries of human ingenuity and drive market dynamics.
Oscar Pulido welcomes Tony Kim, head of the BlackRock Fundamental Equities Global Technology Team, and Michael Gates, lead portfolio manager of BlackRock's target allocation models. Fresh from their interactions with technology leaders in San Francisco and Silicon Valley, Tony and Michael share their insights on the rapid advancements in AI, the efficiencies it brings to the economy, and the promising investment opportunities it unveils across various sectors.
AI, AI Investing, Tech Investing, Technology Investing, Silicon Valley, Tech,
Sources: AI Scaling Laws and Market Structure, Anton Korinek, Professor of Economics, University of Virginia; Bloomberg data as of June 31st 2025; The Complex Truth About AI Computing and Value, Wall Street Journal, March 2nd 2025; Analysis based on BlackRock Global Technology evaluations and calculations; BlackRock and World Bank Group Data, World Bank estimates GDP around $111 trillion, Capital expenditure on AI infrastructure is estimated at around $400 bn.
Written disclosures in each podcast platform and each episode description:
This content is for informational purposes only and is not an offer or a solicitation. Reliance upon information in this material is at the sole discretion of the listener.
Reference to the names of each company mentioned in this communication is merely 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/corporate/compliance/bid-disclosures
<<THEME MUSIC>>
<<TRANSCRIPT>>
Oscar Pulido: Artificial intelligence continues to claim the spotlight and test the limits of human ingenuity. And even after a brief setback early in the year, it continues to drive stock markets. As the technology evolves, so do the investment opportunities. So, where do the greatest prospects reside now and what are the risks?
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.
Today I'm joined by Tony Kim, head of the BlackRock Fundamental Equities Global Technology Team, and Michael Gates, lead portfolio manager of BlackRock's target allocation models. Tony and Michael recently spent time with technology leaders across San Francisco and Silicon Valley, and today they reflect on the rapid pace of progress, the efficiencies AI is creating across the economy and the investment opportunities it is revealing across various sectors of the market.
Tony and Michael, thank you much for joining us on The Bid.
Tony Kim: Thank you. It's a pleasure to be here.
Michael Gates: Thanks, Oscar.
Oscar Pulido: Tony, this is one of our favorite episodes to do over the course of a year because this is where we talk about that tech tour where you take more than 30 BlackRock colleagues around Silicon Valley and San Francisco. In fact, I think you travel more than 300 miles over the course of five days to meet with the leaders of approximately 25 technology companies. So, it sounds like a busy week, and this is the 12th annual tech tour that you've hosted. When we talked about this tour last year, you mentioned that AI was really not much of a topic of discussion for the first nine years of that tour, but that that had started to change and that it became a very prominent topic of the tour. There's obviously been a lot of development in the AI space since we last spoke. So, tell us about those conversations that you had with industry leaders. How has it changed over the past year?
Tony Kim: Yeah, last year it was all about AI, but it was chatbots, the origins of agents. And this year, it was all about AI, but it was AI, what I would call at the frontier, at the edge of what is possible. Really, it's around super intelligence. So, we keep ratcheting up the capabilities of what these AI systems can do and we were looking at AI companies in the full stack, from everything from the infrastructure and the compute to the models, to the applications and now the extensions of AI into the physical world. It's still all about AI, but the scope has expanded materially.
Oscar Pulido: And that term ‘AI at the frontier’ does make it sound like a lot has changed in the past year and that it is more pervasive across different applications and the use case. One of the things that we did this year, again, Tony, is that The Bid team joined you on your tech tour and we were able to get some clips from some of the industry leaders that you met with. And so, on that topic of the rapid pace of change that we're seeing in AI, we want to hear from Jitendra Mohan, CEO and co-founder of Astera Labs.
Jitendra Mohan: The rate of change for AI is insane. It is the fastest pace of innovation that I've seen in my lifetime, arguably in anybody's lifetime. If you look at a couple of years ago, we had the ChatGPT movement and ChatGPT was impressive and magical, but it was still a novelty at the time. And look at where we are today: AI is writing code, AI is finding bugs, these models are able to do reasoning. And I'm pretty sure that there is a virtual AI agent or a physical AI robot in our near future.
Oscar Pulido: Tony, last year when we spoke we also talked about this concept of the technology stack. This is how you described sort of the investment opportunity set and you talked about these as three different layers of investment opportunities. At the bottom you said are the chips and the infrastructure, the middle stack is the AI intelligence layer, and then the top stack are the AI models, the software applications themselves. There's been a lot of, again, attention and focus on the infrastructure — that bottom foundational layer. So, things like semiconductors and data centers is what comes to mind. Is this still the primary investment opportunity or are the investment opportunities now branching out to some other stacks within that, in that foundation?
Tony Kim: This is a great question. I think this is the best way to interrogate AI. If we just look at these three layers: the compute and infrastructure, the intelligence and the models, and then the applications and services.
And so at the bottom of this stack, just the top spenders of physical CapEx is nearly half a trillion dollars this year. It will grow to trillion or more in the coming years. Just to put in context, the global GDP is $100 trillion, so just AI infrastructure is already 0.5%. This is just pure chip and data center investment, is already at half a percent of global GDP. This is the largest investment in human history. We're just in the early stages at that layer, and this continues. And the reason this continues, these are AI factories effectively, they then birth the intelligence. It is created based on how much compute you have. And this intelligence layer they are growing at exponential capabilities - 10 x or more of improvements per year tied to how much compute investment. The world has realized that if you have more compute, we get more intelligence. And so, this is this inexorable linkage between the compute layer and the intelligence layer, and that is giving birth to ever increasingly capable AI systems.
And so, those two are getting the lion's share of the investment, and we're seeing growth that we've never seen before. Companies like OpenAI and Anthropic are growing multiples faster than even Google at the dawn of the internet age. So, we are seeing, as Jitendra mentioned, growth that we've never seen before and the rate of change.
And then at the top layer is what I call the applications layer. And before I was really thinking about the software applications, and I think this is going to change dramatically because the AI will be writing the applications. But I would add one more piece. These AI systems will not just be software applications, but it will be merging software and labor and services together. And so, what we thought were these distinct industries, the capabilities of these AI systems will subsume into the application layer and the service layer. And the service layer is over roughly half or more of the global GDP. And so now you're starting to see the scope and scale of what is potentially coming with these advanced AI systems — it is expanding to bigger and bigger parts of the global GDP.
Oscar Pulido: That was the word that was going through my mind is the scale you mentioned $500 billion in terms of the investment, what it represents as a percentage of global GDP. You touched on some of the companies these days that are growing very fast and how that compares to technology companies of a few decades ago, and just the rate of change being so quick.
We want to hear from another one of the executives that you encountered on this tech tour. This is Sasan Goodarzi, the CEO of Intuit, who sees AI as a tool that could allow customers more time to focus on their real passions.
Sasan Goodarzi: Are there big hard problems that we can solve in ways where a customer never has to lift a finger in running their business, but do all of their finger lifting in what they're passionate about. And that's why I'm so excited about what's possible with AI. I believe it will ignite global innovation in ways that we could never imagine possible, and I can't wait for it to power prosperity in ways that we could never imagine.
Oscar Pulido: So Michael, I want to bring you into the discussion. You were on this tech tour. You're also an active manager. You're managing portfolios and having to make decisions about what to own and what not to own. As someone who invests so broadly across asset classes and sectors, why is it important for you to have a view on technology?
Michael Gates: Well, let's level set a little bit about what's been happening fundamentally for the kinds of companies that are in the public indexes. US Tech over the last 15 years, sales growth rate has been 8% per year compound annual growth rate for the last 15 years, since 2010. If you look at the US market overall, the sales growth rate has been 5%. If you look outside the US, you see growth rates between positive 0.5% and negative 0.5%. Within the US, the IT sector has delivered 50% better earnings growth over the last 15 years at a compound annual rate.
So the ability to beat and raise is something we look for, and where we're finding it is in tech. And within tech where we're finding it is in those companies that are exposed to the AI theme.
Oscar Pulido: Tony, I'd like to come back to you on this topic of AI at the frontier and how much it's evolving. There's a lot of discussion around how AI may supplement human labor in the future, from large language models to all the way to things like humanoids. And so what is your view on this, are we going to see humanoids play a role in our lives anytime soon?
Tony Kim: AI at the frontier, I love this terminology. What we're seeing is this rate of change, an unprecedented rate of change. And humans we're used to linear kinds of change. Growth at 5%, 10%. And what's happening here with AI, we are seeing change at the chip level 2x a year, and we're seeing change at the model intelligence layer 10x a year. So that's like a 20x potential improvement in capabilities, if you take the chip and the model improvements — could be 10x, 20x — but we're seeing this exponential scaling of capability.
So, if we see this kind of non-linear exponential change, and then you compound that one year, two years, three years, four years, you are soon in the thousands of X improvement in just a few years. So, if this rate of change continues between now and let's say 2030, which is just four to five years away, we could see thousands of X capability improvement. So, if we believe that to be the case, and this is AI at the frontier, it is hard to imagine what that means for society. And because the capabilities are improving at this kind of rate, absolutely, yes, you could take the embodiment of this capability and put it into a physical form of a humanoid robot.
Today, I would characterize humanoid robot as the capabilities of a three- or four-year-old child. In two years, it could be a teenager, but in five years we could have very capable, functioning humanoid systems. So, I would say if this is the question on humanoid a little bit, there are the two forms. It's the brain, which is that intelligence capability. And then the physical system of the robot, the mechanical parts. You'll see dramatic improvements on both sides. And so, I think you could get to a point where you could really see that happen before the end of this decade.
Oscar Pulido: And I think the point you mentioned about we're used to linear change and AI really is an example of non-linear change. It's exponential change and it's probably why we spend so much time talking about it as a mega force and a structural driver of returns in the future.
Michael, to come back to you in terms of then the investment opportunity there, there's so much nuance to this space. There's a lot of different winners and losers potentially to come out of the AI revolution. We've seen the dominance of the Magnificent 7 in the US market in particular. But again, how do you think about the role of active management when it comes to this AI theme?
Michael Gates: I keep my eye on the ball in terms of what, is, are the most important indications of financial and economic health. And one of the things we've learned, is that looking at the process of earnings and sales estimate revisions is a really good way to figure out how's it going.
You often get this comment of, oh gosh, the US market's so concentrated, the big companies are such a big part of the index. And that's true, they are a big part. But you might ask the question, is this market concentrated enough? Turn it on its head. And I'll tell you what, if you look at the estimate revisions for the broad index or sectors, what you often find is that there's a subset of companies that are generating really great positive surprises on earnings and sales, and then getting the subsequent positive revisions to their earnings expectations. And that tends to be the largest companies.
Tony Kim: If I were to add to Michael's comments, he is expressing one of my core tenets and observations that I've noticed about the tech sector, what I call a power law. It's not a law, it's an observation. The big get bigger, the winners win, as long as there is tailwind behind these companies. And so, you do see underestimation of the winners just continue to win and win big unless there is some sort of structural change into their competitive position.
And that's what's also fueling these big winners in AI is because there's another element here of AI, which is very unusual, it requires so much capital to play the game.
Oscar Pulido: It's also fair to say, I think we're talking about the concentration of a few companies that have really dominated, but AI as a theme can also then bleed into sectors outside of technology.
And speaking of some of these other industries, Tony, last year when we spoke to you about your tech tour, you you brought up the topic of quantum computing, which maybe was not on the tip of everybody's tongue the way AI seems to be in the last couple of years. And in July, you actually spoke on an investor panel at the Global Quantum Forum in Chicago where they dubbed 2025 as the year of quantum. So you've been an early champion of this technology. What do people need to know about quantum computing, its relationship to AI, and where are we in this journey?
Tony Kim: Yeah, absolutely, great question. Classical computing, which is the Intel, Nvidia, this has been the hallmark of the technology industry for 50 plus years. It just so happens we are now at the dawn of a second computing platform, which was classical computing CPUs and GPUs for 50 years, and that continues. And all of this AI is still based on, as we know, the Nvidia GPU kind of architecture. But these are classical computing architectures for the digital world. I have the opportunity in my career to potentially invest in the second computing platform that will sit adjacent to classical computing, which is quantum computing. Like AI, like fusion, like many other things, these are long, long in development, but we are now at the potential dawn of a utility scale quantum computer.
And what does this do? To just put it simply, it is a completely different computing system that basically computes nature. Nature is based on quantum mechanics at a subatomic level. This is a different kind of computer that can compute the natural world at the subatomic level, which is very different than the digital AI, AGI, classical systems that were driving us to AGI. But this new computing system will sit adjacent to AI systems and it will solve different kinds of problems that we could never solve before with a classical computer. And the output of this new computer is profoundly different and insights to nature, chemistry, biology, drugs, materials, encryption that we've never been able to calculate before, will create new kinds of data.
And then that not only unlocks new discoveries, but it also creates data to help train the AGI system to become even more intelligent. So there is this symbiotic relationship; they don't compete with each other. It just, it's a very different kind of computer. A computer for nature versus the computer of the digital world. And we are at the dawn of this. You mentioned earlier to me about humanoids by 2030, which I think will happen. I think we'll also have quantum computing come onto the scene in a major way also between now and 2030. So again, these next four to five years, it is going to be an epic time of innovation and breakthrough.
Oscar Pulido: It reinforces the point of, this is not linear change, but something unlike what we've seen in the past. We should touch on that, the fact that AI does come with some concerns and we talked about labor and things like humanoids and what that could disrupt in terms of what we know today, but there's also the issue of privacy and security. So I want to go back to our executives that we met throughout the tech tour. We're going to hear from Michael Sentonas, the President of CrowdStrike, which is on the front lines of the effort to provide security, protect privacy as AI grows increasingly sophisticated, and here's what he had to say.
Mike Sentonas: We now live in a world where we are using more and more AI technology in our everyday lives. We're also seeing the attackers get the same sorts of benefits. So we're in this race effectively to build technology, to defend against attacks, to defend and prevent attackers compromising organizations, and attackers that are motivated for financial gain. And we're using this technology to build better systems, better models to prevent these attacks.
Oscar Pulido: So Tony, how are governments and enterprises addressing this and, where's the investment opportunity in this?
Tony Kim: So, cyber is forever here to stay. Before we were addressing human-derived actors using classical techniques to, to try to attack and breach. So now we introduce a super intelligence or a very advanced intelligence. They can write, the AIs can write code, and you will have, let's say millions and billions of AIs that can then create an inexorable rise of attack vectors. And so not only before were we just combating human bad actors, we are now, they can then weaponize themselves using AI-generated cyber vectors. Again, it's a non-linear exponential increase in attack surfaces.
Yet, on the other hand, we could use the AI to help defend against the AI attacking us, and they will be far in excess of what human attackers will be. And so there will be this constant cat and mouse battle of one upmanship of using AI to attack and using AI to defend. This will never go away. In fact, if anything, it becomes even more of an imperative. And so, therefore, the question about governments and companies and I think they're all grappling with this issue — this arms race of weaponizing these AI systems for attack and for defense. And I think that is the direction of travel for the cyber industry.
Oscar Pulido: Michael, this was your first time joining the tech tour. What was your top observation from this marathon week?
Michael Gates: I was really impressed with some of the companies. What was evident to me was was with respect to a couple very important companies in the AI space, they've been anticipating where we are today for years. And they saw that we would be where we are today, which I think many of us find very surprising. They saw this a couple years ago and they are looking to two years ahead. And what they are seeing for two years ahead is really mind blowing.
If you just take it as a fact that they saw what is today a couple years ago and planned accordingly to profit from that development, then you listen to them talk about what is coming in the next couple years and then listen to the indications of what their plans are and the kinds of investments that are happening to make those plans come to being.
Oscar Pulido: And thinking about fast forwarding and what's coming, Tony, if we're having, or I should say, when we have this conversation in September 2026 and we're talking about at that point your 13th annual tech tour, what will we be talking about then in terms of AI and the developments that we've seen at that point?
Tony Kim: I would say we'll have much more conversations, I believe, around this idea of the super intelligence. Are we closer to physical intelligence as an embodiment in physical systems? That will become more of a topic. And then thirdly, I think you mentioned it, societal impact. What is the societal impact and how the face of work changes, how will it change how we work and these ideas, while they're talking about a little bit, I don't know if the mainstream realizes the potential axiom of change that is potentially coming to that part, the societal impact on how we are doing work, and these questions will be, I think, more at the forefront next year.
Oscar Pulido: I already know that's going to be a fascinating episode to listen to as hopefully this one has been as well, just to hear what's going on in the center of the tech world and what's going on with AI. Just listening to the two of you, it feels like so much has changed in the last couple years, but it's also evident there's so much more change coming and we'll look forward to hearing from both of you what that change looks like.
Tony and Michael, thank you so much for sharing your views on AI and the tech tour and thank you for doing it here on The Bid.
Tony Kim: Thanks, Oscar. It's been great to be here.
Michael Gates: Thanks for having me.
Oscar Pulido: If you enjoyed this episode, check out The Intersection of AI and Geopolitics From An Investing Lens where we discuss how rapid advances in AI are influencing global power dynamics. And don't forget to subscribe to The Bid wherever you get your podcasts.
<<THEME MUSIC>>
Spoken disclosures at end of each episode:
This content is for informational purposes only and is not an offer or a solicitation. Reliance upon information in this material is at the sole discretion of the listener.
For full disclosures go to Blackrock.com/corporate/compliance/bid-disclosures
MKTGSH0925U/M-4769469
