Subscribe
Share
Share
Embed
TCW is a leading global asset management firm with over 50 years of investment experience and a broad range of products across fixed income, equities, emerging markets, and alternative investments. In each episode of TCW Investment Perspectives, professionals from the firm share their insights on global trends and events impacting markets and the investment landscape.
Welcome to the TCW Investment Perspectives Podcast, where
our investment professionals share their insights and
expertise on how to make the most of your portfolio.
I'm David Vick, Managing Director and Fixed Income at TCW.
Joining me today is Bo Fifer, Managing Director and Lead Portfolio
Manager of TCW's Global AI Fund, who can talk us through some of the
investment opportunities and some of the news happening in AI today.
Artificial intelligence has rapidly become part of our daily
lives, and it's very clear that there's no going back.
But there's a difference, of course, between using ChatGPT to
help draft a client note or summarize a research article, and
knowing where AI fits into an overall investment portfolio.
So I'm here to join by Bo today to talk through some of these issues.
So welcome to the podcast, Bo.
Yeah, thanks, David. Thanks for having me. Appreciate it. Good to be here.
Great.
So I know that, you know, AI has been around, been popular for a couple
years, two and a half years now or so, into this AI transformation.
Obviously, there's a lot of hype, there's a lot of news about it.
But is it too late to think about adding AI to a portfolio
now, if people miss the boat, or is there more to come?
Well, the short answer is no, they have not missed the boat.
We've actually been investing in AI thematically since 2016.
So it's been around, obviously, in one form or another for a long time.
We are two and a half years into this generative AI boom that
has really caught everybody's attention and imagination.
And we're really looking for ideas now that can change entire industries.
And that can be what we call plain old AI, or it can be generative AI.
Soon we'll have agentic AI and physical AI.
But really, this is, we think, at least a 10-year theme, probably a lot longer.
Early returns are very much in line with prior megacycles in tech.
We went back and we looked at the first five-year performance
of the cloud cycle, the smartphone cycle, the internet
cycle, the PC cycle, and just compared the winners.
We cherry-picked the best-performing stocks in each of those trends.
And through the first two and a half years, you know, the AI
basket is performing very much in line with those prior trends.
So it feels bigger because it's obviously in the news, it's front
page everywhere you turn, people are talking about it everywhere.
So the amount of compute and infrastructure needs to support better performing
or what we call higher IQ, AI, the agentic AI, physical AI, the amount of
compute needed for those things is grossly underappreciated by the market.
We are in this race toward artificial general intelligence, which
is kind of the nebulous term, but it basically just describes
AI that's kind of as smart as a human across a range of topics.
And that race was driven by the hyperscale cloud companies.
They've invested hundreds of billions of
dollars, and that amount is accelerating.
It's increasing.
If you looked at the expected capital expenditures for
2025 today versus what those expectations were a year
ago, they're a couple hundred billion dollars higher.
So the estimates keep going up.
On top of that, we have the tier two clouds.
These are your oracles, your core weeds, your XAI, Stargate coming online.
Those are driving now tens and hundreds of billions of dollars of investment.
And we have sovereign cloud, so nation-state players
who want to localize the data and the AI capability.
Again, tens and hundreds of billions of dollars being committed by then.
And the final driver, we think, will be enterprise.
So as companies figure out how to build AI into their workflows, they
will drive a necessary spend in infrastructure to support those services.
So the early winners have been the core technology or
infrastructure providers, what we would call AI enablers.
That's where I think investor focus should be for now.
So I think that's an interesting segue to, you think
about sort of where those opportunities exist.
I know like on the public fixed income side, the opportunities
you describe in sort of the build out of that compute
power, electricity, data centers, those sorts of things.
From your standpoint, what are the parts of the market you're
looking for, for the best opportunities at this point?
Well, again, our focus is right now on the AI
enablers because this build out is so early.
And I'll give you just one example of that.
When OpenAI released the version of their model that could
create video, they released a couple of data points.
They said 100 million people had signed up for the model,
had used the model over roughly a one-week period, and they
generated several hundred million images, not videos, images.
And that usage, which works out to about one and a half
images per users per day, was enough to, quote unquote,
melt their servers, according to the CEO of the company.
And they ran out, not literally melt, but they
ran out of capacity or supply very quickly.
In our view of the world, we're headed to a place where billions
of people are interacting with AI dozens or hundreds of
times per day, often probably not even being aware of it.
It's just the opportunity cost of not investing in the
low-hanging fruit that is the data center infrastructure.
That's the hurdle we have to clear to invest in some of these other companies.
And there's a lot, and we look for them every day.
I'll give you one example.
Shortly after generative AI came on the scene in late 22, there were a group of
drug companies that held themselves out to be using generative AI to create new
therapies, new drugs that would bind tighter, more specifically to biologic
targets, and therefore create more efficacious treatment with less side effects.
It is and will be a wonderful application of AI.
The drug development process is, in round numbers, a decade.
And there are parts of that process that AI can compress pretty significantly.
But you still have to go through human trials.
There's still a lot.
You have to go through the FDA.
There's just a lot of steps that have to be cleared.
And so those stocks took off.
The first half of 2023 was very good for those companies.
And then I think people realized, wait a minute.
They're still three years away from getting through human trials.
You know, it just was a little bit too early.
And that's kind of been, you can sit down with a clean
sheet of paper and think of so many cool ways to use AI.
The question is, are we there yet, both
from a capability and a cost perspective?
There's a concept that we employ that I think all investors looking to invest
in AI would be wise to think about that we call meaningful and measurable.
So is the impact of AI meaningful to a company's fundamentals?
Is it actually going to move the needle?
And is it measurable?
Is it something that the market can see?
Is it as simple as accelerating revenue growth
or are there KPIs that the company gives out?
Some way for the market to track performance
or you just won't get the credit for it.
When you apply that test outside of the core technology AI enabler group,
it gets pretty hard to find companies who are seeing that impact today.
Well, that's actually a, you know, I think an interesting point to
raise for the next question is, you know, obviously people think about
those big technology names when they talk about investing in AI.
Everybody knows, you know, those companies.
But what's the, what are some of the other things, the most peripheral
ones that I think AI will have an impact on, but we haven't maybe seen?
Or some of those things you're looking at where, if
not this generation of AI, maybe the next one will.
Where are you sort of looking for those sorts of impacts in other sectors?
Yeah, well, again, it's going to be everywhere eventually, right?
But just sticking with healthcare for one second, in this country alone,
we spend four and a half trillion dollars on healthcare every year.
Roughly a third of that we think is administrative.
And so probably a third of that we can target with AI for cost savings.
Potentially, you're talking about taking half a
trillion dollars of cost out of the equation.
That's just a massive number that has economic GDP kind of impact to it.
But more importantly, you know, the AI is getting very good
at detecting disease, cancers, whatever, diagnosing diseases.
It can analyze imaging very well.
If we can find disease earlier, we can treat it, you get better outcomes.
So it will save lives as well.
There's one I've been thinking about lately that I think
is going to be really big, and that's in restaurants.
So I think the restaurant of the future will have digital
waiters, where there's basically a tablet on your table
that you can interact with in conversational English.
And that's how you place your order, get a drink refill, whatever.
You don't have to wait for somebody to come around and get you.
You know, if you need that drink refill, you don't have to flag your
waiter down or a busboy and ask for another iced tea or water or whatever.
You just place your order and it gets sent to the kitchen.
Looking even further out, obviously, that food
can be delivered with physical AI or robots.
That's actually a pretty easy task for them.
Now, that's an industry that, obviously, especially during
the pandemic, they had staffing shortages there, right?
So this is a case where AI probably takes fewer jobs
than it fills jobs that we're not able to fill today.
Otherwise, you know, autonomous vehicles is a very obvious
example of AI in the real world, factory automation.
There's a company that we follow that has built an
AI-based system of robots and statistical software.
They can manage a distribution center with no humans.
It's a lights-out warehouse.
It's super cool.
So like I said, it's coming everywhere.
Those are a couple of examples.
But I don't think you'll be able to hide
from it no matter what your business is.
Yeah, you raised a couple of points that I think are worth following up on.
I know that TCW published a white paper
recently of the impact of AI on labor markets.
And, of course, the widespread fear is that
AI is going to put people out of work.
But I imagine there's a lot more complexity to it than it.
Walk me through your thinking on how AI could
potentially impact labor markets going forward.
Yeah, I think it would be disingenuous for me to
sit here and say AI won't take anybody's job.
Portfolio management jobs are probably the only one that are safe.
But I think, again, there's a lot of places where AI will
fill roles that we're having trouble filling with humans.
But it's like Excel.
If you're old enough like me to remember when Excel first
came out, that was a game changer for worker productivity.
I don't think anybody lost their job because spreadsheets could do thousands
of calculations in a second that we used to have to do on paper by hand.
And AI tools I view is largely the same way.
These are tools that make people more efficient.
A clear example is in coding.
Ironically, AI is very good at writing code.
And coding is one area where we are seeing job destruction.
AI is able to handle a lot of that work.
What I think management teams are going to have a decision to make.
And let's say your base of programmers is suddenly twice as productive
as they were before because you've given them these AI tools.
Your choice will be do I let half of those people go and drop
my incremental profit to the bottom line or do I reinvest that
newfound productivity into my product development roadmap?
Do I want to expand my margins or do I want to pull my pipeline forward?
Got it.
I think those are good points.
So maybe the last question.
This is a tough one.
It's very much crystal ball gazing.
But as you look out, as you look forward, what do you think will
be the biggest impact of AI on the economy as we move forward?
And whether it's, again, this generation of AI or the next one down the road
or where do you think those impacts are going to be the most significant?
Yeah, I don't know if I can put these in chronological order, but
obviously autonomous is going to be huge, whether that's robo taxis
or the autonomous safety functions of self-driving automobiles.
That's a big deal.
Some of the systems, companies have taken two approaches to self-driving cars.
One is rules-based AI.
If you see a red light, come to a stop.
And the other is AI-based self-driving, where no one initially told the car
to stop, but in its training sessions, when it ran through a red light and
got T-boned by a car coming the other way, it got unrewarded for that.
And it learned very quickly that's a bad outcome.
And what's good about the AI approach is the learning curve is extremely steep.
So if you have a fleet of hundreds of thousands or millions of vehicles on
the road that are enabled for self-driving capability, they can send their
experiences up to the cloud and, you know,
the OEM can train its model to get better.
AI is going to have a huge impact on self-driving vehicles of any kind.
Magentic AI is a big one.
This is where AI kind of does tasks for you online, so it can work through web
pages, whether it's booking hotels or flights or just going through your phone.
That's a productivity gain that is going to enhance
our personal lives, but also our professional lives.
And then physical AI, I will say there's some debate around this.
I happen to be a big bull on humanoid robots.
And some of the robots, their physical capability today,
their movements, their balance, the way they interact
with the environment is getting awfully close to human.
And it's pretty easy to put the large language model, you know, the
voice, speech, interactive capability into that device as well.
And we can use that for manufacturing.
We can use that in service industries.
We can use that at home.
A lot of different places we can use that.
And then there's whatever somebody comes up with that nobody's thought of yet.
Those are usually the coolest innovations, the
things that isn't on anybody's radar right now.
But just to kind of bring this back to what we were talking about
at the beginning, the common thread for all of that opportunity
is the need for more compute, for more infrastructure.
It is what is driving all of the gains in artificial intelligence capability.
All of those roads.
We have a lot of levers to pull, but one way or another, they
all revolve around throwing more compute at the problem.
Got it.
Well, thanks for that.
That's all the time we have today.
So thanks, Beau, for joining us.
Let me assure our audience that this is really Beau and me, not our AI avatars.
Maybe next time we'll have that up and ready for you.
For sure.
For more information on TCW strategies, please visit our website at tcw.
com.
Thanks for listening, and we'll pick up next time exploring
more trends and opportunities that are shaping global markets.
Thank you for joining us today on TCW Investment Insights.
For more insights from TCW, please visit tcw.
com slash insights.
This material is for general information purposes only and does not
constitute an offer to sell or solicitation of an offer to buy any security.
TCW, its officers, directors, employees, or clients may have
positions in securities or investments mentioned in this
publication, which positions may change at any time without notice.
While the information and statistical data contained herein are based on sources
believed to be reliable, we do not represent that it is accurate and should
not be relied on as such, or be the basis for an investment decision.
The information contained herein may include preliminary
information and or, quote, forward-looking statements, end quote.
Due to numerous factors, actual events may
differ substantially from those presented.
TCW assumes no duty to update any forward-looking
statements or opinions in this document.
Any opinions expressed herein are current only as of
the time made and are subject to change without notice.
Past performance is no guarantee of future results.