The Bid - AI's 3 Investing Phases
Episode description:
AI has been dominating investing headlines for almost two years, but it's not just a buzzword. It's a powerful technology that's poised to revolutionize industries and economies on a global scale. Investors are asking how will AI reshape job markets, productivity and economic growth? Nicholas Fawcett, a senior economist in the BlackRock Investment Institute joins Oscar to explore what AI means for the broad economy and the different stages of AI's evolution.
Sources: Productivity estimates based on Brynjolfsson, Li, and Raymond (2023), Dell’Acqua et al. (2023) , Cui et al. (2024); Capex spend from BlackRock Investment Institute, Reuters, October 2024; Agricultural data based on IPUMS USA, October 2024
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: 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.
AI has been dominating investing headlines for almost two years, but it's not just a buzzword. It's a powerful technology that's poised to revolutionize industries and economies on a global scale. Investors are asking how will AI reshape job markets, productivity and economic growth? And more importantly, who stands to reap the benefits of this technological revolution.
Joining me is Nicholas Fawcett, a senior economist in the BlackRock Investment Institute. Together we'll explore what AI means for the broad economy and the different stages of AI's evolution. Nicholas will provide an overview of the possible changes across sectors and the strategies investors can use to position themselves for success in this new landscape.
Nicholas, thank you so much for joining us on The Bid.
Nicholas Fawcett: Thanks for having me back, Oscar.
Oscar Pulido: So, Nicholas, we've been talking about AI or artificial intelligence for some time. The BlackRock Investment Institute has, labeled it as one of the mega forces, which means it's a big structural driver of returns. We've talked about the history of AI on The Bid, we've looked into the investible themes that it creates. But curious, how do you see AI playing out?
Nicholas Fawcett: Absolutely right. It's burst into global consciousness a couple of years ago when Chat GPT came out. One of the reasons it's captivated so many people is that AI is trying to embody and then amplify or scale human intelligence. That's why it's so different from the regular ICT boom that we saw from the 1970s and the 1990s, that put a desktop computer on our desks and then a phone in our pockets.
It's the general technology that could rival some of the most important breakthroughs of the past couple of hundred years, like steam power, electrification. It's something that can be embedded in a whole range of different applications and avenues that could transform some of these things in the future.
But there's a lot that we don't know. So, we have to be really humble in, every step of the way because they're so complex, these AI models, and we don’t know how much more complex they're going to become. We don’t know how much energy they're going to need. And we don't know really at this stage what effect they might have on the economy. Is it going to transform a lot of sectors broadly? Is it going to be a more narrow impact? The world could look very different depending on the answer to these questions.
So, in order to get in the game, we think we need to have a good BAT. I'm not talking about sports, I'm talking about a framework for laying out how we see this evolving. And the BAT in question is three things, three phases if you like.
First is build-out, the second is adoption, and the third is transformation. The first phase is where we are now, the build-out phase where we're seeing, in a sense, a race to build-out AI infrastructure - data centers, the chips that are needed and so on.
The second adoption phase is really to come. It's as AI matures and firms start to accelerate their adoption of AI in their regular processes, that's going to require a lot of investment from those firms. And then third is the most interesting long-term impact is transformation.
This is where firms really could unlock the value of ai and there's a story to tell about the economics of all of that there's an important question of what it means for investors because the opportunities are here now, even if some of the kind of longer term structural changes are going to play out over decades.
Oscar Pulido: And so, you mentioned the build-out phase, which is the first of those the three phases that you talked about in BAT, as you cleverly labeled that in the build-out phase. What we have tended to see is a handful of stocks that have really dominated market performance. I'm thinking about companies like Nvidia and the Magnificent seven, top companies in the S&P 500. Is there room to go in terms of the performance of some of those companies?
Nicholas Fawcett: We think there is. You mentioned Nvidia, that's captured the imagination of the tech community because the key provider of all of these chips, underlying AI models, we don't think that rules out further growth. And fundamentally, we're only at the dawn of the AI era companies are racing to build-out ever more complex models that requires a lot more computing power than even we have at the moment. That's going to continue as the technology improves.
What we've noticed is that the underlying GPT models have become an order of magnitude more intelligent with each evolution. So chat GPT3, 4, and in the future 5, 6, 7. And each time you jump up an order of magnitude, you need a lot more computing power. And you can map out with reasonable confidence that the demand is going to be there for the foreseeable future at least. And in a sense, for the firms that are investing in this some of the large tech firms, it's a question of survival rather than should you invest at the margin in data centers or chips or so on, because those who don't compete are risking losing out all of their market share in terms of this opportunity. And the scale that's necessary in order to invest in AI requires you to be a really big player in the first place because the CapEx is so expensive, then you need to continue leveraging that scale in order to continue investing.
It's very difficult for new entrants to get into that market right now because you need a big balance sheet, and that's what these firms have. There's a lot of investor, questions that we hear about whether AI is in a boom bust cycle right now. in a sense, that misses the bigger picture.
History tells us that transformations always create winners and losers. But we think that because AI is potentially so transformational, it has a much broader potential to create quite a lot of value across a whole swathe of the economy and ignoring that could be really missing out on a lot of the gains.
Oscar Pulido: And you touched on CapEx or capital expenditures and the fact that these big tech companies are the big investors in the AI infrastructure. You mentioned they have big balance sheets, and this reminds me of something that Tony Kim also talked about when we talked to him earlier this year about this theme, which is that while the amounts that companies are spending are quite staggering, they have the cash to be able to do it. So, he thinks it's going to continue. How much more will this continue, this capital expenditure boom that we've seen? Where do you see that going?
Nicholas Fawcett: In terms of long-term projection So that's a really big number. Underpinning that is this point that tech needs, computing power needs of AI models are so much bigger than anything that we've seen before. The internet or the search engine technology, that we've seen over the last couple of decades is a lot less computer processing power heavy than AI models. chip costs just directly are going to be a big part of the increasing CapEx spend that we see.
And it's not just that, in fact, it's also the cost of data centers, these chips run like physically really hot, so you need more powerful air conditioning. You need a lot of energy in order to supply the computers. There's a whole ecosystem around the underlying chip that has to be built out too.
In fact, as Tony talked about when he was last, with you, the energy needs are potentially so big that energy suppliers may be struggling to keep up. And so in terms of the risks that we would identify, one of the big near term risks is that they simply can't find enough energy in order to supply that. That's where you see some of the large tech companies investing heavily in even nuclear power stations right now in order to guarantee consistency of energy supply over the coming decades.
Oscar Pulido: And what you're pointing out is that while we think of AI as a technology, it has implications for many different sectors and it truly is transformational. It even impacts the energy sector as well and the demand for energy. Let's stick with Tony Kim for a second, Tony's very passionate about this topic and he's talked about the productivity gains that he thinks it brings to the personal assistant market. but when does AI start to deliver productivity gains for the broader economy?
Nicholas Fawcett: It comes back to the fact that AI is about intelligence, and intelligence is embodied in more, less everything we do, it could significantly boost productivity over the long term.
We already saw a bit of that in the ICT revolution. So, from the mid-nineties to mid-two thousands in the US in particular, ICT was responsible for a big bump up in productivity growth. that's meaningful. In the case of ai, there's a lot of uncertainty, the size and the speed of productivity gains, we just don't quite know yet.
It really boils down to two things. First of all, how far does AI improve the efficiency of specific tasks that you would do within a job, and secondly, how broadly across your job does it help you do things? which is quite big. Think about 10 to 30% time saving on being able to do a task. If you adopt that across every corner of the economy, then it could boost growth by one to two percentage points a year.
That's a slightly abstract number, but when, I was last on, we talked about, demographics being a drag on growth. All of the different drags on growth from the mega forces mean that economies can hope to grow, between one and 2% a year, even as it stands. So a one to two percentage point boost from AI is transformational. but the problem is in reality, that's probably a bit too optimistic.
So it's unlikely that every single job is going to be affected in exactly the same way. Some jobs are a lot more amenable to using AI than others. It could be that maybe a fifth of tasks are impacted.
But again, compared to a baseline in which you're growing somewhere between 1 and 2%, that's a meaningful uplift. Part of the answer is focusing on how AI can help us in existing tasks and existing jobs. some of the really exciting potential is for AI to create jobs that we don't know exist yet and can't even imagine existing.
That's not new if we take the advent of, motorcar, for example, people didn't ride around on horses anymore but it also created the motorway service station and motel network, because all of a sudden people could drive everywhere. It creates new jobs. And in fact, even more recently, and more prosaically, the rise of e-commerce, for example, since the mid 2010s has seen a really rapid boom in employment in warehousing and logistics in the US in particular. There's still potential for AI to create new tasks and new jobs, there's a lot of uncertainty, but that's potentially where a lot of the excitement comes from.
Oscar Pulido: And since you touched on the last time you visited us on the podcast and we did talk about demographics. And as I recall, the one of the ways that economies can try and offset that drag on growth from aging populations is through technology. Innovation and maybe some of the productivity gains that can bring. and that's what you're touching on with what AI could bring in the future. Maybe let's stick with that theme about the labor force and how it impacts workers. How does AI then change the makeup of the economy going forward?
Nicholas Fawcett: So I think AI adoption could massively change the makeup because where labor's deployed, what labor does, that could all change. Even if you don't have a massive improvement in productivity on an aggregate scale, individual kind sectors are really going to be at the forefront of all of the change.
I talked about the shifts that we've seen before, during the, advent of the motorcar. we've seen similar shifts before. just to take a case in point, during the industrial revolution, in So the idea that we've seen big structural transformations is not new in history.
Another great stat that I like half of the jobs that exist today didn't exist in 1950. The idea that all of this technological transformation is going to change the makeup is pretty easy to understand and motivate.
What we've seen over the past and what we think is going to happen is the nature of our jobs is going to change so that instead of doing things alone, we do things with AI as a tool, it's not going to view everyone out of a job. It's just going to make, people more productive, within jobs, even though. Some sectors are going to change quite a bit. So, what it boils down to again, is this question of will AI create new tasks, new jobs, that we can't conceive of right now? And we think that there's a lot of scope for that to happen. And so the makeup is going to change, but it's not going to mean that, we can just do away with labor.
Oscar Pulido: We've talked about the build-out phase, right? that first phase of the AI revolution that you're describing and how it has benefited certain companies in terms of their stock market performance. As we continue along this journey with ai, what are the industries, what are the sectors that you're seeing that are going to benefit? Is it still the ones that have been winning, or does it start to broaden out a bit?
Nicholas Fawcett: Yeah, it's a great question. So this is where we come back to bat. So build-out adoption and transformation. We're still in this build-out phase, and so we still see more room for these build-out related firms, to, to benefit.
And in particular, we talked about some of the chip producers. These are complex chips. They're still going to be made by similar kind of companies that we're seeing at the moment, and we think that it's difficult to get into that market right now. So, there's still ample demand there. we also see an opportunity in the data center and physical architecture around all of these chips.
So, the upstream industries related to that, like industrials, energy, and so on. As we move into the adoption phase, that's where we start to see things broadening out. So some of the companies that can start employing AI to improve their production, and some businesses are already starting to do that.
We're seeing some firms use AI to scan earnings transcripts, so they can assimilate the collect. Assessment of companies at the moment. some people are using it for marketing kind of core centers as a replacement for shrinking workforces and also a lot in the tech industry around coding.
So you can use AI to help you code, even if you know the language already it just massively increases your productivity and those early adopters present opportunities. Now, of course. further out, in this third phase of transformation, we think that there is going to be a wide range of applications, that firms are going to build and that's going to emerge.
And really the focus is going to shift on the industries that have the most potential, the most, AI exposed. Tasks that we can try and find. And the sectors I would pick out there are things like finance, retail, education, healthcare. All areas where there are a lot of jobs that people currently do on their own that could be improved with ai, but not in a way that would do them out of jobs necessarily.
That would just make existing labor more productive. And so you could dramatically increase. I. The potential for those, industries. And we're already seeing it push the boundaries of human knowledge. So there's a widely reported example of using AI tools to try and predict the way proteins fold.
And you can use that to try and then work out how to create new medicines. That's pretty groundbreaking and we're only in the early days and already there are, a lot of examples where there's no way that humans could have been able to make this much progress just on their own.
Oscar Pulido: If you're an investor and you're thinking about the investment landscape and everything that you've touched on, how does somebody best position themselves to benefit from this megaforce?
Nicholas Fawcett: Yeah. And absolutely right. That's where the back comes in again. you don't need to wait is the most important message.
You don't need to wait for the productivity improvements to come in the real economy. Eventually, you can look around and see investment opportunities now. So it's more concentrated right now on, the build-out phase. Even with the rapid evolution of some of the top players in the sheer volume of information, that we're getting and, the limited visibility that we have in some of the emerging tech means that you need to be pretty well informed in order to make a good investment decision.
That's why we think deep expertise is really important. Especially in the environment where there could be a couple of really big winners who win big.
And so relying on broad benchmarks doesn't really fit the bill in this case. in a way it's a kind of existential point. If we don't know what the world could look like in decades to come, the benchmark of, two- or three-decades time is going to look completely different to the benchmark of today anyway.
So what it means is you need to go active, you need more granular insights into, your investment approach and try and identify what the most promising opportunities are. you also have to think about firms that haven't yet floated firms that are still private. 'cause that could be, if you get them right at the early stage, that could be where some of the biggest value gains are to be had. Knowing how to go into that requires a lot of expertise. So that's why we'd say we favor being active and taking an active investment approach, because that's the way to navigate all of the uncertainty and the expertise that you need in order to make the right decision.
Oscar Pulido: Right, be active. And this is a space that is going to continue to evolve for many years. And you've introduced us a new acronym bat, which we will, now internalize and keep in mind build-out, adoption and transformation. And maybe we'll have you back at some point, Nicholas, to see where we are along that journey. Thank you for, sharing this information and thank you for doing it here on The Bid.
Nicholas Fawcett: Thanks for having me. See you again soon.
Oscar Pulido: Thanks for listening to The Bid. If you've enjoyed this conversation, check out my episode with Tony Kim, tech and AI investing trends, a stock pickers take where we discuss the evolving conversation and emerging opportunities for investing in the AI theme.
<<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
MKTGSH1124U/M-3999705
