PM Chris Davis on Investing in the Age of AI

• Back to Video Library

• Disclaimers
• Prospectuses and Forms

Transcript

Chris Davis:

Hi everyone. I want to welcome you to this webcast where we want to take on maybe one of the most central issues facing the market and the economy at this time, which is this revolutionary technology of artificial intelligence. And we wanted to lay out a framework for navigating this tumultuous time. Now, we start with the premise that AI really is perhaps the most consequential technology shift, certainly in modern history and possibly in human history. 

Now, technological revolutions rarely unfold in a smooth, straight line. 
Instead, there are periods of acceleration and deceleration, but importantly, as investors, there are also periods of hype and disappointment. And I think without any question now, we are in the hype period of the cycle. And in this way, we always bear in mind what we call Amara's Law, that new technologies are often overestimated in the short term, but underestimated in the long term.

Now, as investors, what we recognize is that this technology and the adoption of it is going to create radically different outcomes, both by company and by sector. 

Companies that are able to adopt to this cycle will have huge advantages and maybe even broadening competitive advantages relative to their peers, where others are going to be left behind. And to try to create a framework, we've thought of five categories of companies. And what we found as we look at our research through this AI lens, most companies fall into one of these five categories.

What are they? Well, let's start with the category that gets the most hype, which is trying to identify the emerging new winners. Think back to what Microsoft and Intel were in the age of the PC. The emerging winners.

The second category are the enablers. What are those companies enabling this technological revolution?

The third category is who will be the users, who will be able to implement this technology in ways that advantage them relative to their competition.

The fourth category are what we might call insulated or protected companies. Companies that can be, in a sense, almost indifferent to this change.

And at last, and perhaps most important as people position their portfolios, is to understand that there's a category of companies that we call the Walking Dead. Companies whose businesses and fundamental business models are going to be dramatically disrupted.

Now, this is one of the risks in a period like this of passive and momentum investing, 
because both of those approaches are backward-looking, and they have a hard time adapting when there's a dramatic change. So we actually think that many of the fundamental tailwinds that have served passive and momentum-based strategies have come to an end. The wheels are beginning to come off, and we are now seeing an enormous opportunity for investors that can be flexible. They do their research. They have a fundamental value discipline, and are able to position their portfolios actively, not just to take advantage on offense, but also to avoid the big losers in this time of change.

Now let me take a step backwards and look at our investment philosophy and why we think it's suited to this particular time. After all, it goes back more than 50 years. So we've seen revolutions unfold over time, and it's a discipline that rests on fundamental research that is focused on the durability, the quality, and the long-term prospects of each investment. And because of that lens, of course, we have a filter that focuses specifically on the threats and opportunities created by AI.

As we look backwards at past periods of technological revolution and acceleration, they could provide a useful context. If you think about the print or maritime area, the scientific revolution, the industrial revolution, the atomic, and then the digital age, the internet era, all of these can provide context, but the advent of AI represents something that is also fundamentally different and will play out differently. And it will play out across the political landscape, the social landscape, economic, and of course the investment landscape.

Now, the rate of this transformation is going to be neither as fast as the promoters are promising, nor as slow as the deniers would wish. And as I said earlier, we think it's going to follow the pattern of Amara's Law, which simply states that these revolutionary technologies get overestimated and overhyped in the short term, but then underestimated in the long term. So when we think about the overestimating and hype phase where we sit today, we think this is where caution, thoughtfulness, and a broad framework are going to be important safeguards.

So our research indicates that AI is already accelerating the competitive advantage of well-positioned companies. 

It's doing so both by helping their revenue generation, 
but also by reducing costs. And we think we're in the early stage of that cost reduction stage. You put those two things together and that is productivity acceleration.

Now, at the same time, businesses that were once bulletproof,  business models that could have been carved in stone are gradually but relentlessly being undermined. And when we look at strategies that are backward-looking, this type of transition creates enormous risk, right? So that passive momentum approach is much more vulnerable when you have this type of change happening in the underlying economic fundamentals. 

So when we think about the accepted strategies that in our view are taking real risk in this fast-changing environment, well, I've mentioned passive, right? By its nature, it adapts too slowly when the underlying changes are fast. Strategies that are overly diversified, right?
In other words, they own everything, but in times of disruption, that can be an enormous negative. I think one of the most important characteristics that we think is risky in this environment is illiquidity. 

Strategies where the investor money is locked up, where it's unable to move and adapt to changes. Strategies that are reliant on past patterns, and strategies that lack price discipline. 

All of those we think are taking risk in a time of big transition. And our experience is that strategies that are optimized to the past rarely succeed in times of transformation, revolution, or transition. Instead, these times tend to reward active, adaptable managers that are able to use research, stock selection, and maintain a valuation discipline to adapt and modify their approach as the information begins to unfold.

Now, before diving into the company-specific implications of AI, we want to develop and share with you a general framework that balances its transformative power with the uncertainty and the fact that this technology continues to unfold on a daily, weekly, monthly basis. Now, what we start with as a fundamental supposition is that AI is one of the most powerful tools that mankind has yet developed,  and that multiple companies and multiple countries are developing this tool and the technologies around it simultaneously. Now, because of that, it seems unlikely thus far that there will be a winner-take-all model, that it would be characterized by the sort of network effects that you see in certain technologies. So when you think of network effects, think of the invention of the telephone where a single network garnered almost all of the economic advantage of that.

Information utilities like Google, social networks like Facebook. At this point, we don't see that winner-take-all framework unfolding yet. We're keeping our eyes open, but our working assumption is that it's not a winner-take-all business as of yet. Instead, we think intelligence will be widely available almost as a utility, and therefore that the economics may accrue to the users rather than the creators. 

Now, in this way, think about the printing press, think about railroads or the interstate highway system. Think about broadband maybe as more useful reference than the network effects, the winner take all effects of things like telephony. Now, like electricity, AI will increase productivity and efficiency. 

That's why we think so much about the users, and that will lead to an economy-wide shift that is going to favor early adapters. 

Now, when we think about some of the downsides that we can think of, the disruption of white collar knowledge workers, I think it's going to be analogous to what blue collar workers experienced during the time of factory automation and globalization, just that factories could accomplish more with fewer workers and the employment implications of that.

So we would expect employment to adjust over time with a particular pressure being put on the white collar knowledge workers, what Elon Musk once called the laptop class.

Think of the fact that 80% of the US labor force used to work on farms versus 2% today, but the key will be the pace of that change. So the speed of the AI rollout really does make short-term unemployment shocks something that should be on investors' screens. 

We're beginning to see the layoff announcements, and we think over time that will continue, particularly in these highly compensated areas of the economy.

Now, as with the atomic age, we also have to be aware of the risks that are created by the misuse of this powerful, powerful tool. So the misuse that can come from bad actors, it can be individuals, it can be terrorist organizations, it can be rogue states. So certainly there's a background of higher risk and a higher shock that we need to be aware of. 

We saw an awareness of these risks just recently with Anthropic's rollout of their newest technology and the recognition of how quickly it could be used to make financial institutions or software vulnerable to hackers.

Now turning from this general framework to the really important issue of stock selection, AI is going to impact businesses in different ways. 

And understanding how it will affect business models has to be central to investment analysis. This is not new for us. We think back over the last 60 years about the amount of technological disruption, whether it was the internet, advertising, undermining newspaper business models, or whether it was the replacement of the mainframe with the PC, these sorts of technological disruptions have occurred in the past and an important part of long-term investment success is being early in recognizing the transformative powers of new technologies. 

Now, as I mentioned earlier, we classify all of the companies we look at into five categories based on the potential impact of AI on their business. But importantly, there is going to be enormous dispersion both within and between these categories. 

So stock-specific analysis will matter more than ever.

Now, as I mentioned earlier, we group each of the investments we look at when we put them through this lens into one of five categories. 

Those categories are the Emerging Winners. Who will emerge dominant as a result of this new technology? Who are the Enablers of this technology? Who are the companies that will be able to use it best to improve their competitive positioning? What companies are insulated from these changes? In other words, that in a sense can function as business as usual, and then most importantly, who are the Walking Dead? Who are the companies whose competitive position is going to be eroded and their business model is going to dissipate? Now, understanding these categories will help investors navigate these increasingly volatile market conditions.

Now the first category and the category where pundits and promoters spend most of their time is what we call the Emerging Winners. What companies could emerge as dominant as this transformation unfolds? Now, it's important to take a step back and recognize that when you're in the early stages of these new technologies, that identifying the winners can be enormously risky. 

In fact, there can be broad consensus about who those winners will be that's proven wildly wrong. For example, if you look back at the early stages of the automobile revolution, there were more than 3,000 American auto companies, of which only three survived. 

Now, if you look back at the early days of the internet, in March of 2000, there were 371 publicly traded internet companies. The vast majority of these were wiped out, many went bankrupt, and the favorite winners at the time, the surefire low-risk blue chip internet companies, which were Yahoo, AOL, and Cisco, are all shadows of what they once were.

So it's a dangerous time in this uncertain world to try to put chips down on winners. The key is, even if you identify a durable business, buying it at a time of maximum hype is enormously risky. 

So you have both a stock selection risk and a valuation risk. If you look back again at that March of 2000 period, even the companies that proved to be tremendously durable, think of Amazon and Booking and so on, you were able to buy at significantly lower prices as the period of disillusionment unfolded. 

And of course, many of the winners of the internet were not even in investors' minds in March of 2000. Google and Facebook, for example, were to come public several years later. So at this early stage, we think it's important to avoid companies that have little or no earnings because they're reliant on capital markets to continue to fund their growth. If they have unproven business models, if they haven't been able to monetize their technology, and if they have insatiable capital needs,  we think that you could have winners emerge from that, but it's enormously risky. The fact that we have so much uncertainty around pricing power, that at the moment switching costs are quite low, these increase the risk in these capital-intensive business models. And importantly, we think when there are aggressive growth assumptions baked into valuations, it can turn even wonderful companies into mediocre investments. 

When we think, for example, about some of the hype and valuation enthusiasm that's baked into the companies that design chips used for AI, for example, we notice that they assume that there are no viable competitors that emerge. In other words, they project out very high margins that you would have much more associated with a monopolistic company, and they do that for long periods of time into the future. And yet when we look at the companies, we see that large customers are very busy at work designing proprietary chips to compete. So having your customers trying to compete with you can increase risk, and such valuations simply do not leave a margin of safety. So, as a result, we favor companies when we think about trying to identify emerging winners that have characteristics like these, companies that have the raw material necessary to power AI. So, think about strong, wide data sets, proprietary data.

We think about companies that already have enormous cash flow that they can use to fund their growth. They don't have to keep going to creditors or to capital markets.  We like proven business models, we like reasonable valuations,  and we like companies where the use of the AI will simultaneously be enhancing their core cash flow.

So think about things like some of the largest social media companies, companies where the use of AI are making their ad platforms more valuable.

The next category that we focus on are what we call the Enablers. These are companies whose businesses will benefit from the enormous levels of capital spending to fund AI,  but they will benefit whether or not the returns on that spending end up being satisfactory or poor.

So in this way, because they provide materials and equipment, think of them as analogous to the companies, the picks and shovels during the gold rush. 

They are benefiting from the speculative capital pouring in without taking the same risk. So we think of these particularly in industries like the chip manufacturers, right? This global oligopoly of the companies that have the wherewithal and the knowledge to manufacture these complex chips, the companies that provide equipment to those factories, right?

The semiconductor capital equipment providers, or the companies that make the chips that are not considered nearly as glamorous, but are absolutely necessary for translating data from the real world, the analog world, into the digital world. So think about analog chip designers and chip manufacturers. Now, those as enablers seem quite straightforward, but there's a second category of enablers, which people are beginning to spend more and more attention noticing, which are those companies that are able to provide the energy and the raw materials that will be necessary to fuel these enormous power consumptive applications of AI. So here, think about natural gas, companies with long reserves that are well located and the increased demand for that natural gas as a fuel. Or think about things like companies that have long-lived reserves of copper because the transmission of power and the electrification, the increasing demand of electricity, what we call the electrification of the economy, will consume a great deal of copper as well.

As we think about the third category, we look back to innovations and revolutions like the transportation network, railroads, or broadband. And what we recognize is enormous amounts of the value went not to the companies that built that infrastructure, but to the companies that used the infrastructure. So we call this category the users of AI. Now, here, the industries that are data-intensive,  such as financial services, health insurance, that rely on expensive knowledge workers, here's where we could see the greatest acceleration of productivity. Remember, productivity is output per man-hour. By having this new tool for knowledge workers, we expect a great surge in output per man-hour, the amount of work each knowledge worker can do. And that surge in productivity drives profitability. So AI is going to create opportunities to retool these business models, both to reduce costs, and to improve profitability, and drive the top line by creating more value for customers.

So this creates a lot of disruption in this knowledge worker base, in the laptop class, and it also requires large investments. And therefore, the companies in these industries that are able to make the investments, the companies that have the modern tech infrastructure, you need the right tech stack to be able to apply AI across an enterprise, and you need tech-savvy managements. Most companies are going to move too slowly. 

They have the classic innovators dilemma. They have a good business, earning good profits. They don't want to make the big investments, retool their system, go through layoffs, and they'll just move too slowly. And that will really advantage the companies that are able to make the investments and will become competitively advantaged and will be able to gather market share and improve profitability. So the Users of AI is, I think, one of the categories we focus the most on as we see this tool rolling out through the economy.

So now we move from the Emerging Winners, the Enablers, the Users, to a category that is almost the opposite. And the idea of this category, which we call the Insulated or the protected, came from a comment that Jeff Bezos made some years ago, when he said people are always asking me what's going to change. And sometimes an equally important question is, what is not going to change? 

So in a world where there's so much of a spotlight on AI and what's changing, there's a very little attention being paid to companies that are not in the center of this feeding frenzy. And these are the ones we think of as potentially more valuable, because their business models are protected.

Now, these companies often grow more slowly. Sometimes they have some cyclicality in their earnings or economic sensitivity, and so they really require valuation sensitivity, but these companies have been ignored and many of them now trade at attractive valuations and attractive valuations on somewhat depressed or non-peak earnings.

So here we think of some food suppliers. We think of some generic drug manufacturers, resort operators, medical supply manufacturers. None of these are glamorous high-growth industries, but to be able to buy durable companies that are insulated from some of these revolutionary changes at really attractive valuations, that can play an important stabilizing role in an investor's portfolio.

The final category we think about as fundamental research-driven investors and that is critical to avoid is a category we call the Walking Dead. These are companies that had once durable business models, business models that could have been carved in stone, but that are completely undermined by the advent of the new technology. And we think looking in the next 10 years, that investment returns will be built as much by avoiding or getting out of losers as it will be by picking winners. 

And research can really help identify these declines before they've happened. Now, let me give you a real example of this. When you look back at the late '90s and early 2000s, Kodak was a completely dominant company, and digital cameras were beginning to roll out. Well, here's an amazing thing to think about. By 2001, 30 million digital cameras had been sold.

That meant that there were 30 million people in the world who knew firsthand that film was dead, that the idea of buying a roll of film and putting it in your camera and taking your 36 pictures and then taking it out and driving to the store and dropping it off and going home and then driving again to pick up the photos, only then to see that your kids' eyes were closed. That once you had a digital camera, it was absolutely clear how much more useful it was, how much cheaper it was, how much more convenient it was. 30 million cameras had been sold, and yet Kodak was still in the top third of the S&P 500. So it's clear to see that the ability to be able to move in front of the indexes and these passive strategies gradually catching up is a critical and key advantage that we expect active investors to have in the next 10 years.

Local newspapers were another example. It was clear that people could go onto jobs.com or onto Monster or into cars.com and immediately have a better experience at almost no cost. And yet classified ads were still 50 to 75% of the profit stream of local newspapers that maintained huge status in the indexes and in long-term investor portfolios. So active investors had time to exit, while passive investors had to go down with the ship.

And now, those were ones where you had a clear shift that was in the business model, but things will also happen in very unexpected ways. And a useful heuristic to think about is the saying that when the iPhone was launched, flashlight makers were not nervous. So there will be also unexpected transformations, and many of which will happen quickly. The ability of research-driven active investors to get out first, while the index, while they sort of go down gradually and go down with the ship, I think that's going to be a huge advantage, and it's going to be a great disadvantage to rely on backward-looking models.

Companies that have been reliable dividend payers, reliable growers, low-risk strategies can be enormously disrupted, and that can happen quite quickly. 

If you were to think back only five or seven years ago and try to create a safe portfolio, you would easily put in companies like Nike and Starbucks and Estee Lauder and Diageo and Brown-Forman, for example. And of course you would've had suffered catastrophic losses because the world had changed. And we think this is one of the reasons that active management has been so well positioned in recent years.

It's the ability to move in front of these sorts of changes. Well certainly, recognizing the admonition that it's always better to praise by name and criticize by category, I think simply as you look at companies whose business models are based on things that can now be done virtually for free, by virtually anyone, you can see that there are a number of white collar service firms that have a lot of this vulnerability. We've also seen it in some of the SaaS companies, Software as a Service companies that were beloved darlings just a short time ago.

So we think one of the reasons that active investors with our discipline are so well-positioned in this time is that the indexes and the passive investors and the backward-looking strategies have to react so much more slowly that the research that we're able to do can be more valuable than ever.

Maybe the simplest phrasing of the advantage that research-driven active management can have in this fast-changing environment was given by Kenny Rogers in the song, The Gambler, when he said, "You got to know when to walk away and know when to run." And again, avoiding losers in this time of transition, avoiding this Walking Dead category is going to be enormously important to investor returns and those models that rely on backward-looking data instead of forward-looking research are going to be enormously disadvantaged. 

Before wrapping up this idea of the Walking Dead, I want to highlight one additional area of enormous risk that's been confusing for us to look at, which is the illiquidity risk that people have been willing to take in pension plans, endowments, and increasingly as individuals by locking up their money and strategies that have five, seven, or 10-year lockups, private equity, private credit, this sort of thing, because the ability to change your mind in a time of transformation in the economy is enormously valuable. And conversely, the inability to change your mind can be enormously punitive. So we think it's a dangerous time for people to be taking liquidity risk. And of course, we're beginning to see that unfold as you read some of the headlines about people trying to get their money out and unable to do so.

Putting this all together, the advent of AI represents a unique moment in human history, and it is going to transform the global economic landscape. And what we know is that periods of disruption reward flexibility, reward rigorous research, it rewards a valuation discipline, and active decision-making. Now, while this rate of this transformation is uncertain, the shift is already underway. We are seeing it in the companies that we're studying. And during this time of change, investors have to value flexibility, right? 

Having a lack of flexibility is, to use Charlie Munger's colorful phrase, it's like being a one-legged man in an ass-kicking contest. Passive strategies, overly diversified approaches, approaches that rely on backward-looking data, such and such a company has always paid a dividend, so they always will. Illiquid strategies, strategies where you can't change your mind, where they can't get out and reposition. All of these strategies are enormously disadvantaged in this fast-changing world, and instead we think investors should be shifting away from richly-valued indexes.

They should be shifting away from illiquid strategies, and they should be shifting towards active, adaptable, and flexible managers, managers who can incorporate both the threats,  but also the opportunities that are posed by this powerful new technology, into their research discipline, into their stock selection, and into their portfolios. Thank you very much.

Related Videos