Full episode description:
AI investing continues to shape markets as artificial intelligence (AI) moves beyond software and into the physical infrastructure of the global economy. From data centers and chips to space-based compute, autonomous trucks and humanoid robotics, the AI buildout is creating new questions about scarcity, supply chains and where value may accrue next.
In this episode of The Bid, host Oscar Pulido is joined by Tony Kim, Head of the Global Technology Team within BlackRock Fundamental Equities. Fresh from his 13th annual technology tour across San Francisco and Silicon Valley, Tony shares what he heard from leading innovators and how the AI conversation has evolved from model development to compute, infrastructure, physical AI and the changing shape of the technology stack.
Tony explains why AI investing may increasingly require looking across multiple layers of the ecosystem: the physical layer of power, chips, data centers and cloud infrastructure; the intelligence layer of foundation models; and the application and services layer where disruption remains a central question. The discussion also explores how AI is creating both scarcity and abundance, why data center demand is reshaping supply chains, and how countries and companies tied to the compute build-out may be positioned differently from more service-oriented parts of the market.
Key insights include:
• How AI investing is expanding from model development into space, robotics and physical systems.
• Why data centers, chips, power and cloud infrastructure form the physical foundation of the AI stack.
• Where supply constraints in energy, materials, optics, packaging and semiconductors may shape the next phase of AI growth.
• How low Earth orbit satellites could create new forms of AI data and, potentially, new compute architectures.
• Why autonomous vehicles, self-driving trucks and humanoid robots are part of the broader physical AI story.
• How the shift in market value toward compute and model-centric companies is reshaping stock market trends.
Keywords: AI investing, artificial intelligence, technology investing, capital markets, megaforces, data centers, robotics, stock market trends
Source: BlackRock Fundamental Equities analysis of AI-related capex spending through 2030, as of July 2026; 'How much does a GW of data center capacity actually cost' Investing.com, 2025; S&P Global Indices as at July 14th 2026
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 any company or investment strategy mentioned is for illustrative purposes only and not investment advice. In the UK and non-European Economic Area countries, this is authorized and regulated by the Financial Conduct Authority. In the European Economic Area, this is authorized and regulated by the Netherlands Authority for the Financial Markets. For full disclosures, visit blackrock.com/corporate/compliance/bid-disclosures.
Oscar Pulido: Artificial intelligence continues to write the market narrative and defy even the loftiest expectations as a powerful driver of corporate earnings. At the same time, AI disruption has led to some stock market air pockets this year, with software at the epicenter. But beneath these headlines, the AI juggernaut powers on at an unprecedented pace, begging the question: where are the opportunities 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 global technology team within BlackRock Fundamental Equities. Tony is just back from his 13th annual tech tour across San Francisco and Silicon Valley. He brings fresh reflections on the ingenuity and optimism he's hearing directly from AI innovators and discusses the evolving investment opportunities he sees being created as AI not only crosses but creates new frontiers.
Tony, thank you so much for joining us on The Bid.
Tony Kim: It's awesome to be here.
Oscar Pulido: Well, Tony, this is one of my favorite episodes of the year. This is the episode where we talk about your annual tech tour. We started doing this a couple of years ago. You just recently hosted your 13th annual tech tour in June where you, and 35 BlackRock colleagues, boarded a bus and traveled more than 300 miles to meet with some of the world's most innovative technology companies in San Francisco and Silicon Valley.
I'm happy to say that this year I was actually on that bus. I haven't been able to join in the past, but this year I was able to make it, and it was really eye-opening to see just the access that you get to some of these leading companies. Tell us what struck you as the biggest change in the conversations with the companies this year relative to prior years?
Tony Kim: Really, this is a snapshot. It's a marker of what is the state of play in tech. And every year it changes. In the last three years it's obviously been about AI. What struck me about this year versus, let's say, last two years is, the AI wave continues, but it's starting to expand and branch out, right? The size and scope of these AI projects are bigger than ever, than anything that we maybe contemplated two years ago.
Secondly, the data centers, the design of those chips, the data centers, is also changing because the requirements are ever-expanding. And so, you see, the forward planning of these completely new designs, and then they're looking very different than what we thought two years ago.
Another one is, around, the models themselves are changing. More and more they're becoming more like the human brain and that's driving a whole design change and ethos in design. Those AIs are also being very tightly coupled with the compute and you're seeing this co-design, tight integration. You're seeing the expansion of many more types of models. And we're talking about world models and physical AI systems which precipitate a different kind of AI computation and AI model.
The other thing that's we've seen is a changing of the guard, if you will, of the kind of technology companies that are really aligned and benefiting with AI. And clearly it's moved to a compute and model-centric type of technology company, and it's raised questions about other types of technology companies prior to AI. And those questions are still being asked, and they're asked with more veracity about their business models and their role in a post-AGI world.
I would say those are the topics of this year's tour that were maybe, they were somewhat in the background, but they much more to the forefront right now.
Oscar Pulido: And Tony, just listening to your response, yeah, AI was obviously the persistent theme across all these conversations. And I remember over the last couple of years when we've talked about your tech tour, you had mentioned AI had been picking up in terms of prevalence in the conversation, but now it's basically the entire conversation.
One of the things that I always enjoy about this episode and the tech tour is that we get to hear from some of the industry leaders at these companies that you met with. Let's listen to Ashley Johnson, President and CFO of Planet Labs, offering a peek into this world…
Ashley Johnson: The pace of innovation is almost breathtaking right now. Our teams have brought the Silicon Valley way of doing software to the space industry. So, as you see changes coming into the iPhone through the camera capabilities, the chip capabilities, the communications infrastructure, we can harness that in every launch of satellites that we do.
And on the data side, we're able to take all of the advancements that are happening in AI and use that to build out solutions for our customers so that they can understand what's happening around them - from climate change through to movements of people and migration, all the way through to geopolitical and military dynamics - almost in real time and certainly daily. The pace of innovation is allowing all of these customers to get so much more data from space and use that to help life on Earth, which is the mission of our company.
Oscar Pulido: So Tony, the conversation about AI is evolving so quickly that now, as Ashley points out, we're talking about AI and the space economy. And, when we talk about AI in space, we're not talking about terrestrial cement data centers, but we're talking about clusters of satellites that push the boundaries of what we're able to do on planet Earth. Help me understand and help our listeners understand, what does this idea of AI in space actually look like?
Tony Kim: Planet Labs is a satellite maker, a provider of low Earth orbit satellites. And what Ashley's talking about there is, I think, two things. First of all, the models today are predominantly language models, voice, text, human text, human language. And if you have a constellation in space, it's around low Earth orbit, which they do, can you also build another kind of AI model around, satellite image data? And so you are building another kind of modality of capability. So that's one. And then there can be a real-time nature to it, where it's constantly being updated - a new kind of data.
So, they're leveraging that constellation and the data of that satellite image. Then the other point I think she's alluding to - and other companies like SpaceX and Blue Origin and others are talking about - can we also then leverage that same kind of constellation for compute?
So one is another kind of data to help build new kinds of models, and the other one is can we leverage the terrestrial, the burdens of terrestrial computing, which we all know about, the energy, the power, the regulations, the permitting, and then move it into space. And then if you can move some of that compute into space, then we can have huge constellations of compute clusters that can, basically effectively, beam inference tokens down to earth.
Oscar Pulido: I'm thinking about the BII mid-year outlook, we recently spoke to Jean Boivin, and he talked about this concept of scarcity, that the buildout of AI creates scarcity in certain materials. There's a scarcity of energy. But by moving some of this into space, you actually start to address some of that scarcity potentially when it comes to things like energy. One of the things that you also talked about on the tour this year as like a new observation was more discussions around physical AI, things like humanoids and autonomous vehicles.
Let's hear from Lior Ron, COO of Waabi, who shared his excitement over self-driving trucks.
Lior Ron: This year and next year is finally all about real adoptions. Not pilots, not theory. Self-driving trucks on the road without a driver beginning of next year in Texas. So, this is a fundamental sort of sea change in the industry. We've all been waiting for this moment for over a decade. It's finally upon us.
Oscar Pulido: So Tony, Lior says it's finally upon us. When do you expect to see real adoption of AI in the physical space?
Tony Kim: So, I would say, broadly these next five years, we are on the precipice of adoption. In some cases, mass adoption. I think we are already seeing that with cars. trucks are very close behind. And then humanoids and robotics will follow.
But I'd say in the next five years could we see a million plus cars and humanoids in production in operation? I think that's a possibility. And I think it all leverages the work that's been done on the AI model side and the compute. You needed to have all of that in place, and now we're bringing that intelligence in the embodiment physical systems. The core technology built around building these models, language models, world models, self-driving models sitting on a foundation of compute. And it's getting better and better.
And then once those capabilities get to a point, you can then embody them into physical things, be it a car, a truck, a tank, a satellite, a submarine, a robot. So, I think this is all, the next five years. And so you will see more and more companies that start to scale this activity. And we're just in the very early days of that.
Oscar Pulido: And Tony, I mentioned this before, but as exciting as all these developments are, there are supply constraints when it comes to building out the infrastructure to support all of this progress. What are some of those constraints when you think about all that it takes to rewire the global economy to be more AI-centric?
Tony Kim: Yeah, I like the term rewiring. We need to build an entirely new internet. The internet pre-AI was built on, on a very basic foundation of chips, very little power, small data centers, CPUs, a little bit of memory. Today's data centers and tomorrow's data centers are an order of magnitude more compute, order of magnitude more energy, order of magnitude bigger.
And these are radically different things. It's almost like an alien data center of today versus yesterday. And tomorrow's data center is going to even drive the physics. Another common theme of the trip was, we're hitting laws of physics in many things, which is precipitating changes in a data center architecturally to change yet again to fit the demands of AI. And so, when you do these step functions, order of magnitude type of changes, it creates what you also alluded to before, scarcity or abundance. And in particular, we're suffering from scarcity because we have not built our entire ecosystem to build to this scale and to build to this design.
These are new designs, never before contemplated, and the designs are so much larger than we've ever contemplated. For example, the data center spending that we are contemplating, forecast for the next five years will be probably over $10 trillion these next five years. And if you think about what does that mean? Let's equate that to gigawatts of energy. So oftentimes compute is equated to a computing unit of energy. and so roughly today to build a gigawatt is $40 to $50 billion. I'm sure you've had many guests that would quote that number. Even if we took the lower number, the $40 billion number, and if I said $10 trillion or more of spending, that's 250 gigawatts or more. One gigawatt can roughly serve Manhattan. So, that means we need to build 250 Manhattans worth of compute and energy just for that contemplated capital spending. Think of then what it takes to build 250 Manhattans worth of compute and energy, and that trickles down to the entire ecosystem, the entire supply chain. And we can start with the laws of physics, right? The energy itself, the grid, the power, all the materials, the copper, and then you go into the chip world, the substrates, the packaging, the semi equipment tools, the wafer, the foundry, the optics.
Everything is short. Everything is constrained. Everything is scarce. Not everything, but almost most things. And then you throw at that, these 250 Manhattan-equivalent demand over the next five years, and this is the situation we're in. That is the world of scarcity.
Oscar Pulido: And you're right. There was a lot of mention of physics and just science in general when I was sitting in some of these company meetings. I quickly had to search up some of the terminology to understand what was being discussed.
But the themes were around there's not enough power. How do you transmit that power? Where do you get the basic raw materials? How do you cool the chip that is running at a very high heat?
One of the executives who talked about this was Jim Anderson, the CEO of Coherent. Let's listen to him speak about the demand being created by AI and his company's focus on scaling supply.
Jim Anderson: So, the demand is very strong, both the near-term demand signals from our customers, but also the long-term demand, is incredibly strong. And so, the biggest challenge is scaling supply as fast as possible. We're a manufacturing company, almost everything we sell, we make. And so, to meet the demand, we've got to scale the supply, and that's what we're focused on 100% of the time.
Oscar Pulido: Jim reiterates this point about the importance of supply right now and supply constraints. And so, taking all of this into consideration, you've talked about the buildout layer of the AI stack, and maybe you can elaborate a little bit more on what that stack is. Things like the application layers, robotics, and we've talked about space, but where is your team seeing the investment opportunities right now?
Tony Kim: There are three main areas, and then within each area there's some sub-layers. At the very base layer of the stack is what I call the physical layer of AI, the physics that we often talked about on this tour. The energy, the power, the chips themselves, and then the data centers and the cloud. Those are all physical manifestations of AI. And, and quite frankly, in these last three or four years of AI, most of the value has been accrued there.
The second layer is the intelligence layer. This never existed before AI. And here you have these foundation model labs, building AGI and lots of, people building models. That's that second layer, and we've seen a lot of value being created there as well. Companies building models, either the existing public companies and we have all these new companies. And by the way, at that bottom layer of compute, most of those are existing public companies. And so, we're seeing a lot of value also accrue to the model layer.
And then at the top layer, the third main category, there's a few areas here, but that is really what I would call the applications and services. Now, this is software, these are services, and when you look at that layer, there's more questions being asked. You've seen the SaaS-pocalypse concerns, around all these various service industries also being, being disrupted. That's been a difficult area of the stack in the last three years.
And, in fact, when you look at these three layers, there has been trillions of dollars of value that have shifted from that top layer down to the model and the compute layer.
And then if you look at over the last three years, those bottom two layers have materially outperformed and added much more market cap than that top layer. And that's because, the nature of AI is fundamentally about building that intelligence, and to build the intelligence and to run the intelligence is you need to compute.
And that is in effect, in my opinion, has been pulling value away from that top layer of the stack, that in the stock markets around the world, Asian countries, Taiwan, Japan, Korea, in the last two years have done amazingly, materially outperforming the U.S. market. And why is that? Because those economies are aligned to this part of this bottom layer of the stack. They're building, and particularly that supply chain. Everything that we're talking about, the constraints, the scarcity, they're building the wafers, they're building the substrates, they're building the robots, they're building the power equipment, the semiconductor, the wafers. That whole world of compute is aligned to Japan, Korea, Taiwan. The U.S. economy, it's mostly a service economy. And I don't know where we sit today, but S&P is up 5% to 10% year to date. So that's radically been less than these Asian economies, because they're aligned much more, their entire economies are aligned to the build-out of this compute layer and this infrastructure layer.
And then within that S&P, you have that service software layer that's negative year to date, and then the compute complex, is up materially year to date. And, so the market is also shaking. It is rejiggering, as you said, the rebuild, the rejiggering, and reassessment of where value lies in this AI stack.
Oscar Pulido: Right. In other words, which companies are going to adopt AI more quickly and then perhaps gain more share versus their competitors, and perhaps that leads a little bit more to this dispersion that you see in the performance in stock prices. Tony, when we talk about AI and we think about the future, one of the storylines is that AI will create a lot of abundance. That could be innovation, could be economic growth. Of course, there are some people, though, who think about the disruption that it provides to just the day-to-day life for humans and the tasks that they perform and the role that they play in the economy.
So, as we wrap up and think about the future, we also spoke with Mala Tejwani, the COO of World Labs, who offered a balanced reflection on AI
Mala Tejwani: Working at the forefront in AI means the change of pace isn't monthly or quarterly. It's weekly and daily. My advice is as fast as the world is changing, as innovative as it's becoming, as much as there are tasks that you no longer have to do, don't outsource thinking and thought and how you're problem-solving. Because those are the pieces that humans will have to continue to do, both being creative as well as being thoughtful and analytical.
Oscar Pulido: So Tony, you've been an AI optimist for years now, and were talking about it before any of us were talking about it, but what do you see as some of the risks, either to the investment story or maybe some of the risks that are created by AI that people aren't fully appreciating?
Tony Kim: Mala is a remarkable person if you look into her backstory. And, so I just want to say that, first of all. Secondly, the clear risks are, as we know, these AIs are becoming more intelligent and cognitive labor as we know it, people are questioning and worried about that. Now, what I would say is, I study a lot of history and I look back throughout time. The current composition of the world is two-thirds, 70% are services.
We're a service economy, a service global economy. But that wasn't always the case. Back in ancient Rome, we were 95% agriculture, very little services. Then we hit the Industrial Revolution, and then gradually that kind of spawned a shift from the agriculture to industrial goods and manufacturing and the beginnings of the service economy.
We then hit the information age, and then we became, radically, a service economy. So, as we go forward, when I look in the past, there has always been disruption. But imagine going forward, yes, many of the jobs we have today likely will be transitioned but imagine if you think of the art of the possible. What is the art of the possible? And imagine if each of us had 100-200 digital employees that we can each use, and wield to our will, what could you do?
What if we had capabilities with AI and compute to just assign them to tasks to try to figure out very intractable problems in life sciences, in energy, in materials, in physics itself. We could solve very different kinds of problems and that could unlock new capabilities. What I do know is that history has shown us that the economy, the shape of the economy, the composition of the economy has changed, and I suspect it will change again but we don't focus enough on what is the art of the possible. And, and that's why I remain an AI optimist.
Oscar Pulido: And some people know, and you mentioned it, that you are a historian in addition to being a technology investor. So, I think your point is that the economy has gone through periods of disruption in the past, but it has adapted, and that we should keep our eyes on the risks, but also on the opportunities that AI could bring.
Tony, I mentioned that I joined you on the tour this year. I have to confess, it was only for one day. You did five days worth of touring. I was impressed with the stamina that you and the team have to visit so many companies across San Francisco and Silicon Valley. But again, it just goes to show how close you are to this theme and the companies that are at the forefront of this change. Thanks for doing that tour, and thanks for coming to The Bid and sharing the stories with us.
Tony Kim: Oscar, it's always a pleasure. Because your presence, it was the best tour ever.
Oscar Pulido: Thanks for listening to this episode of The Bid. If you wanna check out the previous episodes on Tony's tech tour in Silicon Valley, check out the links in the show notes. Up next, we'll be taking a look back on the biggest trends of the year to date. So, if you've missed a couple of episodes, don't worry. We'll catch you up. Subscribe to The Bid and don't miss the episode
Disclosure
This material is intended for information purposes only, and does not constitute investment advice, a recommendation or an offer or solicitation to purchase or sell any securities, funds or strategies to any person in any jurisdiction in which an offer, solicitation, purchase or sale would be unlawful under the securities laws of such jurisdiction. The opinions expressed are as of the date of publication and are subject to change without notice. Reliance upon information in this material is at the sole discretion of the reader. Investing involves risks. BlackRock does and may seek to do business with companies covered in this podcast. As a result, readers should be aware that the firm may have a conflict of interest that could affect the objectivity of this podcast.
For more information go https://www.blackrock.com/corporate/compliance/bid-disclosures
For investors in Israel: BlackRock Investment Management (UK) Limited is not licensed under Israel’s Regulation of Investment Advice, Investment Marketing and Portfolio Management Law, 5755-1995 (the 'Advice Law'), nor does it carry insurance thereunder.
In the UK and Non-European Economic Area (EEA) countries: this is Issued by BlackRock Investment Management (UK) Limited, authorised and regulated by the Financial Conduct Authority. Registered office: 12 Throgmorton Avenue, London, EC2N 2DL. Tel:+ 44 (0)20 7743 3000. Registered in England and Wales No. 02020394. For your protection telephone calls are usually recorded. Please refer to the Financial Conduct Authority website for a list of authorised activities conducted by BlackRock.
In the European Economic Area (EEA): this is Issued by BlackRock (Netherlands) B.V. is authorised and regulated by the Netherlands Authority for the Financial Markets. Registered office Amstelplein 1, 1096 HA, Amsterdam, Tel: 020 – 549 5200, Tel: 31-20- 549-5200. Trade Register No. 17068311 For your protection telephone calls are usually recorded.
©2026 BlackRock, Inc. All Rights Reserved. BLACKROCK is a registered trademark of BlackRock, Inc. All other trademarks are those of their respective owners.
MKTG0726-5748882-EXP0727
