How should I be positioned? with Matt McLennan (First Eagle) and Jason Draho (UBS CIO)
The recent dialogue between Matt McLennan and Jason Draho highlights a cautious outlook on the US economy, influenced by evolving monetary policy and the emerging AI capital expenditure cycle. Per the full note , McLennan emphasizes the importance of strategic portfolio positioning amid potential market risks, suggesting that the current momentum in equity markets may not be sustainable without continued Fed support. This conversation can shape traders' expectations on FX movements, particularly if macroeconomic indicators signal a shift in economic dynamics or Fed policy adjustments.
What the desk is arguing
The desk emphasizes the need for careful positioning as the US macroeconomic landscape evolves and the potential impact of Fed monetary policy looms. McLennan’s insights regarding portfolio risk management are underscored as vital in times of volatility, highlighting that recent equity gains may be fragile. Per the full note , heightened scrutiny on AI investments also introduces an interesting dynamic in capital flows that FX traders should consider.
Moreover, the desk is particularly focused on upcoming inflation indicators, which could influence the Fed’s next moves on interest rates. The debate over the sustainability of recent gains in equity markets, tied closely to Fed support, could lead to volatility that FX pairs traditionally react to, especially USD exposures.
Where it sits in our coverage
As for our current FX positioning strategy, we see the USD/EUR pair's consensus target at 1.075, with a corresponding range of 1.04 to 1.12. Specifically, jpmorgan targets 1.10 for March 2026, while bofa sits lower at 1.04 in the same tenor.
This desk's cautious perspective aligns with jpmorgan, which reflects similar apprehensions about sustainability in the current market environment. However, it diverges from bofa, which appears more bearish on the USD's trajectory against the EUR, suggesting a potential downgrade in USD performance against the Euro in the near-term outlook.
01The conversation between McLennan and Draho highlights a cautious outlook on equity market sustainability.
02Focus on Fed monetary policy signals will be crucial for future currency positioning.
03Investment in AI sectors may alter capital flows and influence FX dynamics.
04The current strategies emphasize risk management in volatile markets.
Market implications
Traders should monitor USD/EUR levels around the consensus target of 1.075, especially as economic data emerges that could prompt a Fed response. Additionally, the next inflation report will be a critical catalyst influencing market sentiment ahead of Fed meetings.
Risks to this view
A shift in monetary policy direction from the Fed back to a tightening stance could significantly reverse the current market outlook, particularly if accompanied by negative economic indicators that could trigger a flight to safe-haven currencies. Additionally, unexpected geopolitical events or a major downturn in AI-capex sentiment could also alter investment flows.
ubs
Hello, everyone. Edwin Marrero here. Welcome back to How Should I Be Positioned on the UBS Market Moves podcast channel.
On this podcast, we like to catch up with our industry colleagues and thought leaders to exchange views on the market and macro environment, along with shared thinking when it comes to asset allocation. Joining me here today at the table for today's episode at the 1285 Podcast Studio in New York. Glad to welcome back from the UBS Chief Investment Office within UBS FSI, the head of asset allocation for the Americas, Jason Dreho.
Joining Jason and I today from our partners at First Eagle is Matt McLennan. Matt is the head of the global value team and portfolio manager at First Eagle. Matt is also a member of the UBS Research Advisory Board, a group of experienced market leaders and industry experts who collaborate with the Chief Investment Office at UBS to explore key themes driving markets, the economy, and long-term investing.
Their perspectives help broaden the conversation around today's most important investment topics and bring valuable insights to UBS clients and advisors. With that, Jason, Matt, great to be at the table with you both. Welcome back.
Matt, welcome to the new podcast studio here at 1285. Thanks so much. Thanks for joining us.
Yes, in our plush new studios here. Oh, they're delightful. With a lot to cover on today's episode, let's get right into it.
So the first question will go to you, Matt. I'd like to start with the cyclical outlook for the U.S. and global economy. How do you see the macro environment playing out over the next year?
When we think about the cyclical outlook for the macro environment, we're always very focused on corporate profits because we view corporate profits as the core kernel of the business cycle. And I think one of the things that's quite striking about the environment right now is that we see very positive trends in corporate profits in the U.S. and not just in the U.S., in the U.K., in Europe, in Japan, in emerging markets. And so across the world, we've seen upward revisions to earnings forecasts, and that's been mirrored in fixed income markets with the yield curve steepening that we've seen.
I guess the one exception to that is China, where corporate profits have been sluggish, and the 10-year bond deal there is well below the developed market average. So the direction of travel, at least as it would relate to corporate profits, is quite constructive right now. The question we have is more so around the margins that are expected as we sort of look forward into 2027 and beyond.
The market is pricing unusually high corporate profit margins compared to history, and I think there's been three outsized drivers of those margins. I think the first and arguably the most important is that we're in a window where we're running very large fiscal deficits relative to the level of unemployment in the economy. And if you think of those deficits in the public sector, that produces income in the private sector.
And if deficits are three or four percentage points wider than they ordinarily would be at this stage of the cycle, that means corporate profits are doing well. Why corporate profits? Well, the second driver is that household savings rates are very low.
They've dipped below 3% in the United States. That means the surplus that's coming from the fiscal stimulus is all going through the corporate profit margins. In essence, the households absorbed the tariff hikes and the energy price hikes that we've seen from the partial closure of the Strait of Hormuz.
And I guess the third thing that we view as having contributed to corporate profitability is almost the accounting timing mismatch from the AI CapEx boom. In a sense, the money that's being spent by the hyperscalers on CapEx is the revenue and profit for the semiconductor industry today, but the expense from that is only going to be depreciated across time for the hyperscalers. So we have this unusual confluence of large fiscal deficits, low household savings rates, and an accounting timing mismatch from the AI CapEx cycle.
And I think the challenge for markets as we sort of look forward into what happens 2027 and beyond is that higher real interest rates, the need for household savings rates to normalize, and the catch up in depreciation to the CapEx run rate of the hyperscalers are going to be weighing against further margin improvement from what's already expected by the market. So that's just something to reflect on. Jason, what did you take away from Matt's thoughts there?
And can you share with us the economic outlook through the balance of 2026 from the CIO vantage point? I think I wrote down seven or eight questions. I'm going to follow up with Matt.
On earnings a lot of really interesting points there. But I'm going to zoom out a little bit and think more about the macro because I think that does influence the cycle, but maybe not. That's one of the questions.
Just from the macro perspective, and just to name drop a little bit, I was on Bloomberg this morning, and I made a comment about the macro being sort of, I think, kind of relatively benign, all things considered. It generated a bit of a guffaw from one of the hosts, but I sort of want to explain why I think that from the year ahead, relatively benign. And that broad level of at least U.S. growth and even GDP, global growth, we're getting like 2% trend.
Some of the data we got from July suggests maybe a little weakness, but when you kind of zoom out the whole year, you strip out some of the accounting aspects of GDP, but we had 3% private sector final demand in the first half of the year for the U.S. economy. Maybe it moderates the two, but it's still a pretty decent kind of economic growth environment. Inflation, certainly still, that's the kind of the story, the macro story.
It is coming down, and you can point to different measures to tell a slightly better story. But ultimately, I think there are enough structural drivers that inflation should continue to moderate. Covid being, we've had multiple supply side shocks, there could be more.
This feels like that's the story of this decade. If this macro plays out, the Fed doesn't have to raise rates, and macro does not become a headwind for the markets. And that's the key kind of thing that I would play.
That's sort of our macro view. So supportive for equities, but I guess the first question I have for you kind of following up is on this earnings, the way you think about the earnings story. And the numbers we've had for Q1 far exceeded expectations going in, Q2 far exceeded, and we're like 90% of through the earnings season as we record this.
At the start of the year, the bottom up analyst consensus for the S&P was around 13-ish percent, 12% to 14% earnings growth for the S&P this year. Right now, given the Q1, Q2 earnings, plus the guidance we had, it's somewhere like around 26%, 27%. And that's stripping out some one-time investment aspects.
So that far exceeds the growth improvement story. So at this point in time is when you look at the profit cycle, has it sort of become disconnected, do you think, from the economic cycle? And does it have to do with the structural story that is AI that is like, you know, something that's like you can't look at the economic cycle necessarily to understand the earnings cycle right now because AI is a structural story?
Look, I think the very high rate of earnings growth relative to nominal GDP reflects some of those points I made before, the unusually large fiscal deficits. I think, you know, we're at a point, and I guess we can delve into this more detail as the conversation goes on, but with real interest rates above 2%, the dynamic there becomes more challenging. So I think we've seen a lot of the benefit from easier fiscal policy already.
The household savings rate went from 6% plus down to 3% minus. That can't really repeat itself. And so that's been a source of margin improvement.
And the AI contribution to earnings power, I think that accounting mismatch is very important. And I think one thing to reflect on there is that the hyperscalers went from spending less than 20% of their cash flow from operations 15, 20 years ago on data center build-out to this year more than 100%. And so that window of CapEx ramp relative to their cash flows is in the rear vision mirror now.
And as we look forward to 2027 and 2028, the rate of growth in CapEx is going to moderate and the benefit from that accounting mismatch is going to moderate. And so it would be our expectation that we may be at the peak of earnings growth rates at this point in time. And I guess the other thing I'd mention is that as earnings and margins have inflected higher the last 18 months, we've started to see the job market bottom.
If we look at the job openings rate, for example, it troughed at the end of last year and has now started to inflect a little higher. And in fact, the job openings rate is higher than the unemployment rate in the economy, which means that we may start to see some wage pressure rebuild in the economy, just as everyone was assuming we'd get all of these layoffs from AI. And so I keep a close eye on the wage growth picture because I think that's often the killer of corporate profit growth.
So picking up on one of the points regarding the earnings, you might check the fiscal stimulus. From a fiscal perspective, there was some stimulus earlier this year due to the one big bill, the consumer side. But when we look at the magnitude of fiscal deficits for this decade, it's hovering around 6% for multiple years.
It's not necessarily a huge delta. One thing I think about what's going on in the marketplace right now is that a lot of focus on the issuance of debt by these hyperscalers and others. And in simple terms, I'm almost thinking what's going on is the hyperscalers who have really good balance sheets are able to issue at really low rates.
They're turning around, taking that money, borrowing it, and then turning around and basically buying and spending it over to semiconductor companies. And so the semiconductor companies are leveraging the balance sheets of the hyperscalers that are AAA to borrow. And that money is like basically every dollar I'm going to have for bondholders, essentially going 90 cents on every dollar is going to the bottom line of earnings.
So some of the earnings stories are just basically, oh, they're just levering up the industry overall. And as that plays out, if that sort of peters out, one that maybe is a risk to earnings because it feels like it's a bit of an inverted pyramid. If there's that marginal borrower doesn't want to provide the money, then suddenly does this all come kind of cascading down?
Or once that ends, do the earnings of those companies drop significantly? So just from that kind of private sector dynamic that's going on, how much of a risk do you think this is? Is how much of this is maybe being inflated by all this borrowing?
And once it slows down, the growth is going to drop off. The market's not priced for that. Yeah, no, I think the markets will have to absorb a deceleration in the coming couple of years because you've already levered balance sheets to a certain extent if they want to stay within their current rating paradigm.
And as we mentioned earlier, you're already spending the cash flow from operations that you had in place on data centers, which as you point out, is the revenues for the semiconductor companies. And I guess the other thing to weigh here is expectations in the marketplace. We talked about the fact that it's relatively benign backdrop, whether it's GDP growth or earnings growth, but the markets are now priced for a benign backdrop.
Credit spreads, if we look at the high yield option adjusted credit spreads, they're pretty much back to where they were in 2007 in the lows of the prior cycle. Equity multiples, even though they've come down from their peaks with earnings growth, are well above their generational averages. And when we just look at household asset allocations, households are more allocated to equities compared to real estate than at any time that we can see since World War II.
And the last couple of times that we approached these levels of equity allocation were the late 60s and the late 90s, which preceded a period where real assets performed a lot better for the decade ahead. And I think it's an interesting moment because we are at that point in time where the digital economy is meeting the physical economy. And you can see that it's not just DRAM prices that are going up.
Copper has made new highs and we sort of see that broadening impact from the CapEx into other sectors, power prices, even insurance premiums for the facilities that have been put up. So we're starting to see that demand seep into other areas of the market. On the point about kind of the accounting mismatch, you know, you're kind of expensing it over time, but you're booking the revenues now.
We go back, it was about a year ago when this started to become the whole set of circularity of financing, like, you know, these suppliers, you know, we'll give you the money, you buy it, then everyone's happy until at some point it doesn't work. And that was what happened in the late 90s, early 2000s with a lot of telecoms companies. When you look at that kind of stuff and these kind of nature of these contracts, you have to obviously make assumptions about the economics of how this will play out.
Does this worry you significantly? Do you think this is, like, some of the concerns get overdone? And if there is sort of, like, real pain points, like, this is, like, what part of that sort of really gets you concerned?
Well, I think right now the demand for compute is very high relative to supply. Like, the vacancy rate in data centers is extremely low. And so, you know, it would appear that there is real demand for this.
But if we think about the run rate of CapEx spend and you project that forward over four to five years, it's a huge amount of incremental capital that's being built. And it sort of begs the question of whether the revenues from inferencing and selling, you know, access to the well-trained frontier models is going to be enough to generate a good return on all of that capital. And that's just an unknown question.
I think there is a little bit of the risk that when we, you know, akin to when we had the railroad booms, you didn't need five railroads connecting the same two cities. You only needed one or two. So we have a lot of money being deployed into competing frontier models.
Maybe we see some rationalization over time. But I think the one area that we see a lot of future compute demand coming from is inferencing. And I guess the way we've addressed this in our portfolio is that we have a kind of a barbell, if you will, in terms of how we're thinking about investing in AI.
On the one hand, we have invested in companies that have an advantaged cost curve position in terms of building frontier models because they might already have billions of existing subscribers. We do have some semiconductor investments, but we're actually quite a bit less than the market. We've focused on some Asian semiconductor manufacturers that are not pricing these margins to continue indefinitely.
But I think what's interesting is in the last couple of months, we've seen quite a bit of opportunity in the software space. And we've really narrowed our search to software companies that are system of record software companies because they control the data that's going to be needed for inferencing. And so an LLM can scrape all the publicly available information, but it can't necessarily access proprietary information.
So we've, I guess, zeroed in on those companies that we think have proprietary data. And we see them providing a lot more demand for compute as customers use it in everything from CAD CAM to payroll and other applications. So this is really good.
I've had a lot of questions recently about kind of AI, like the whole, I call it AI flywheel. It feels like that's, in some sense, the macro story. By flywheel, I mean, you have a situation where the applications go out there, so the enterprises have to use it to find there actually is productivity, enhancing center costs, savings, benefits to justify the inference demand, the demand for compute.
If there's demand for compute, that leads to the investment that takes place. And as long as that's all kind of working together, it's sort of reinforcing the moment that sort of, you know, kind of becomes questionable. Then you wonder, like, does this sort of, again, sort of cascade in a negative way?
This is purely anecdotal, but coming out of the Q2 earnings season, it seems like one of the narratives in the markets are companies of different types, whether it could be software companies, cloud companies, have demonstrated, like, we can monetize this investment in different ways. And the markets are kind of less anxious about that than they would have been maybe three months ago, six months ago. But this ebbs and flows.
Like, this will, I'm sure, will be, like, questions about overinvestment again multiple times before this ultimately resolves itself. Are you on that aspect, on these questions of overinvestment, are you more comfortable with or, like, think about it in a different way? Is it something like there's just so much demand, like, the investment story will continue to play out for, like, we have transparency for a few years, or how do you?
Honestly, I don't know. I mean, you know, when I speak to the companies, they see it continuing to play out. You know, on the other hand, I see real frictions.
And we talked about the increase in DRAM prices. You know, we've seen increases in power prices. We've seen copper hit new highs.
And these frictions at the margin erode the economics for the folks building capacity. And if we see labor costs start to inflect higher, that's going to be an interesting one at a macro level. And I just want to come back to the point before you'd mentioned that we've seen 6% fiscal deficits for the better part of a decade.
It's not that much of an incremental fiscal pulse. If anything, the deficit was slightly down year over year. But the one reason we think fiscal risk is perhaps the least well-perceived risk in the market is that markets are actually starting to price a different kind of risk today.
You know, it's interesting to me that the Fed has been on hold for some time. And in fact, that comes after reducing short rates by 175 basis points. But the 10 and the 30-year bond yield have broken out on the upside after a 40-year downtrend and are making new highs.
And so the long end of the curve is pricing something a little bit different from just the front end of the curve. And I would submit to you that I think that there is a bit of a risk premium creeping into the long end of the curve to, you know, looking at the supply of treasuries. And, you know, I think the key thing is that we added sedimentary layers of sovereign debt in the wake of COVID, basically.
But if I look at the left and the right, there's no political constituency to tackle entitlements. And so that's, you know, continuing to grow apace. With the geopolitical backdrop, defense expenditure is on the rise.
And the government has to roll its debt to the current higher interest rate deck over time. All of these things mean it's very difficult to contain the deficit at these levels on a rolling forward basis. And as an investor, you're looking at that supply coming.
And at the same time, some of the traditional buyers of U.S. government debt have diminished their enthusiasm. Some of the reserve-accumulating countries internationally, after we sanctioned the Russians' ability to access their treasury reserves, behavior changed. And I think what's another price signal that's worth paying attention to here is that in the very short term, you see gold acting like a tip.
As real rates go up, its price goes down and vice versa. But over the last five years, while real value of treasuries has gone down meaningfully, the real price of gold has gone up meaningfully. And so the repricing of gold and the widening of term premia tells me that fiscal risk is actually starting to be priced, and that's why we're focused on it.
So on this point, you know, standard question that I'm sure you would get, that I would get is, well, what level of rates can equities handle? You know, and there's no kind of magic number, but with the 30-year at 5.3%, the 10-year at roughly 4.75%, you know, it's often like it's not so much the level, it could be the pace. But at some point, there's a question.
So how do you, if this is a slow bleed of rates going higher, I mean, I'm not sure if we're going to get a list trust moment where suddenly the 30-year is up 80 basis points in a week. I think that's unlikely, but you never know. But how do you think about what is, you know, what is sort of like a point like, as long as growth is okay, nominal GDP is okay, equities can handle a 10-year at 5%.
I mean, you know. Yeah, I think you're right. I think what would spook equities is a discontinuous move at the long end of the curve, like the list trust moment.
Because the symmetry of wider fiscal deficits than we've had historically, which have helped corporate profit margins would be if you had the bond market vigilantes coming in and demanding a higher yield, it essentially be the bond markets forcing a fiscal consolidation on the United States. And if we had to move from 6% or 7% deficits to 2% or 3% deficits and primary surpluses, that would be very negative for corporate profit margins. And it would be a negative pulse to nominal growth.
And so I think equity markets would react to that kind of shock risk. The thing that's harder to know is just, you know, with a market trading north of 20 times earnings, you can peg your long-term expected return, but it's probably in the 6% to 7% range. So if the long end of the curve starts to creep into that zone, then there's substitution demand and, you know, you're buying an asset that has less than perfect correlation.
The risk, of course, is, you know, the other risk we haven't talked about is the straight of all moves hasn't still really reopened. And so the worst case scenario, and it's not one we're predicting, but it's just one that one has to be mindful of, is that kind of replay of the 1970s type environment where you had structural fiscal deficits, you had energy prices that end up being higher than people expected in real terms. And there was a point at which the growth narrative in the markets just got whacked, basically.
It was the nifty-fifty back in the late 60s, early 70s, but those long-duration growth stories, as the bond market started to price higher for longer, derated very meaningfully in the 1970s. And so that's something that one should be attuned to in their portfolios. And I think, you know, in a way, if I could sort of summarize it, I think that the market's been extremely concentrated, and it embeds expectations for a benign outlook.
And so I think you're being compensated to broaden your asset allocation right now internationally and to other sectors in the U.S. that have been less in favor. So on asset allocation, like that's my responsibility, not just, you know, equities, U.S., ex-U.S., but fixed income, commodities, the whole, everything. And I agree that the, like, rates, rising rates are at risk.
It makes financing the deficit more expensive. I think the number that was floating around last week was like 1.4 trillion annually spent on just on debt servicing costs. And this number will go higher as long as interest rates are higher.
It just, you know, the kind of question, like, do high interest rates beget high interest rates because— Yeah, they do. There's a kind of an autocorrelation or, you know, a self-referencing function. As long as the government's running a big primary deficit before interest expense, as you roll your debt to the next level of interest rates, the deficit goes higher.
And so you're chasing your tail. So this gets into, well, what's the solution? You mentioned, like, you know, you have, you can have politicians cut benefits.
And cut benefits could also entail tax. Like, you're taking money away from people, either, like, their support, like, you know, health care benefits or raising taxes, which is politically unpopular. So what everyone wants, what, like, an economist would say, well, the idea is you grow your way out of it.
That's easier said than done, unless you try and get higher nominal GDP growth, which is, all right, if you're going to have a nominal GDP of 6% and your deficit is 6%, at least your nominal debt to GDP is staying constant. You're buying yourself time for a political solution, which is what I think all politicians, and certainly the Federal Reserve, would be thinking. So if that environment— and we will get a sense of, like, you know, in the coming weeks of, like, what Kevin Warsh is thinking as the speech at Jackson Hole.
You know, you could lay out a case. He's sort of already made a case. He, like, believes in the AI productivity story.
One could conjecture—this is just pure conjecture—that, you know, he would believe, I don't want to kill the golden goose because AI is going to save all of us later on. So we need to let some investment play out, let the economy run a lot, because later on, AI will be this massive disinflationary growth force. And it's going to be the 1990s Greenspan canola over again.
Whether you believe it or not, that's the playbook. One thing is that, well, as an asset allocator, I'd rather own an asset that can benefit from high nominal GDP growth versus not, meaning I'd rather own equities versus bonds. So you mentioned, like, 6% or 7% we may get there, but, like, inflation is high enough, like my real return, I'd rather own equities.
And so I get all the arguments, but if I have to pick the two, it's like, well, it's sort of like, you know, the best house in a bad neighborhood, perhaps. Yeah, and I think you have a choice within equities. So if the concern is you—or the mental challenges that you want to avail yourself of the higher nominal drift rate in the economy, and equities is obviously a good path to do so.
How do you do that, perhaps, in a way that's a little more defensive than just being long, the higher beta sectors of the market, which is really the concentration of many market participants right now. And I think, you know, as we survey the world scene and we look at other regimes that have had higher rates of nominal drift, there are low beta ways to participate in that nominal drift in equity markets. Some of the leading consumer staple companies of the world, for example, that have derated as bond yields have moved higher, but they're really more like tips than they are like fixed nominal securities, because they have pricing power over time.
You know, whether you're, you know, 40% of the world's toothpaste market, or you're the leading brewer, or, you know, you're the leading soap and shampoo company in the world, these companies over time are a lower beta way to participate in that nominal drift, and so perhaps the opportunity becomes to sort of introduce a little prudence into the portfolio through sectors like that. And outside the United States, there are many great consumer staples available with 6% unlevered free cash flow yields. So if you're getting, to your point, mid-single-digit nominal growth and those kinds of free cash flow yields, you have a lower beta way to produce attractive returns.
And I would say similarly in the U.S., healthcare has been one sector that's been very out of favor, that you don't have to just buy the pharmaceutical companies that have portfolios that run off over time. You can buy companies with 40, 50, 60% market share and niches that have derated to similar free cash flow yields. And so I guess part of our portfolio, part of the ballast in our portfolio is low beta equities in healthcare and staples, commercial services with recurring revenue streams, etc.
And so we've broadened. I guess tech was a very big sector waiting for us a decade ago. It's more modest now.
I want to come back to some of the portfolio discussion, but just an input to that is valuations. And you touched on the multiple this year for the S&P 500 on a forward-looking basis has come down. It was roughly 22 early in January.
Now it's 20-ish, give or take. But I've also had an argument that the rating for the S&P should not go back to those power levels because there's so much capex spending. These companies that had been asset-light, they're actually becoming more asset-heavy.
And therefore, you're not willing to pay the same kind of premiums. Interesting argument. I'm not sure.
I haven't thought too deeply about it. But how would you think, as these kind of companies transition to something where they own a lot of physical assets, their multiple should not be as high? Yeah, I think the free cash flow conversion per dollar of earnings has gone down.
And so what we think matters ultimately is what price are you paying relative to free cash flow. So I think there is some merit to that argument that there's more capital drag. And it's interesting when you look at the capex of the hyperscalers, it now dwarfs what the Exxons and the Chevrons of the world have been spending.
And those companies over time traded at lower PE ratios because free cash flow on average was a little less because you had to invest in the capital stock. And so I think there'll be an element of that. But I think the real reason that multiples have to come down is that you're capitalizing margins that are very high relative to history.
At a moment where savings rates are low relative to history, we talked about what had to be done to consolidate the budget, how to do that when household savings rates are 2.7%. And so I think that the combination of low household savings rates, big fiscal deficits, that timing mismatch we've discussed with the AI capex in terms of the expense recognition versus the revenue recognition. I think all of those are logical reasons why I think one could argue that the multiple should be lower than it was when margins were more depressed.
So kind of related to that is if I look at semiconductor companies, look at Korean equities. Their forward multiple is five, which you'd say this is like, you know, super depressed and like maybe it's depressed for a reason. Now it's at that level because their market is dominated by at least two well-known semiconductor companies whose earnings, the numbers are staggering that from 2025 to 2027, their earnings can go from 10 billion to 150 billion.
And the market is just not willing to pay for it because they think this is unsustainable. It's mostly just price increase, the volumes can't increase, and the hyperscalers just have unlimited funds right now to buy this. But as a value investor, I mean, it's not a simple metric.
That's relatively cheap. Some of the semiconductor companies have said, well, we have contracts that will allow us to not be as cyclical. This is multiple years.
Again, if you want people to buy your stock, that's a nice story to tell. But as someone who actually has to think of does this sort of meet the criteria? Is there something fundamentally different about like this part of the tech ecosystem because of this AI build-out that starts to make change how we think about it?
Has that changed? So we had actually made a substantial investment into one of those two stocks that you were alluding to 18 months ago. It doesn't seem that long ago, but at the time, that whole sector was trading below book value, which was asynchronous with what was going on in the world of GPUs.
But the simple insight our tech analyst, Manish, had was that for every GPU, you're going to need a memory stack. And in fact, actually, as you move towards more inferencing and more complex forms of compute, memory becomes increasingly important. And while it's not as nearly monopolistic as the GPU market is, you really have three players who are essentially 90% of the market.
And I think because they've gone through cycles, they're inherently a little bit prudent about the pace at which they add capacity. And so it is our feeling that while we've taken some money off the table, we still have a meaningful position in one of those securities. Because our discussions with the company and the environment around them would suggest that people are locking up longer-term arrangements.
And even on his first conference call, Elon Musk at SpaceX basically said, this is an environment where memory pricing and memory demand doesn't have to just go up by the normal 10 to 15%, it has to go up by order of magnitude. And so I guess we're having to sort of look and see where the new equilibrium settles out. I think the right way to think about those businesses in the markets are being a little skeptical, as you said, mid-single-digit multiples.
But if you look out a few years, they're really being valued at one times what they think book value will be in two or three years' time. So there's not a huge amount of expectation in those stocks relative to their current reality, which is very strong. But I think most people are rightly erring on the side of caution, given history.
So our time is coming close to an end. It goes by fast. I guess the first question I would have goes back to this portfolio discussion.
And you already gave examples of things that you like, whether it's some of the more defensive-oriented staples, things of that sort. It does seem like within equities right now that AI is such a dominant theme. And frankly, it matters for all asset classes to different degrees.
And the concern I have, I think anyone who's doing asset allocation, but even just pure equities, is understanding the various kind of exposures you have to AI. It's not just a couple of stocks anymore. It can impact industrials who are helping to build data centers.
So how do you think about and assess and then maybe diversify away from equity risk? Where is it exposed? How do you understand?
Are you getting compensated for those risks as you think about different opportunities? Yeah. So one of the interesting subcurrents of the market over the last couple of months is that if you looked at the implied volatility of constituents in the S&P 500, it got to a very high ratio relative to the implied volatility of the index.
So with that second-order AI effect, there's been a lot more idiosyncratic divergence between stock prices. The correlation structure between individual securities broke down. And that led to quite a bit of opportunity for us.
If you look at our turnover in our portfolio, it's usually quite low. It's often only about 10% a year, so decade holding period. Our turnover for the last couple of months was almost 2x that, or maybe even a little more, for the very reason that you suggest, that the market is feeling in the dark, trying to price the AI disruption risk.
And I guess the opportunity for us has been to identify businesses that we think have proprietary data, because the market's been discounting all software companies or all services companies in a similar manner, like a common factor discount. But I think some of these businesses will persist and have much more duration than the market thinks. And on the flip side, I think the productivity story for AI, it's wonderful, but our belief is it's going to take a lot longer to get to some form of AGI than markets hope, in the sense that if you look at what the founders of the industry, like Demis, who was at Alphabet, and Yann LeCun at Merrill in their labs business, they basically made the point that these LLMs are essentially word prediction devices.
But real intelligence requires world models or simulations of causal relationships between actual phenomena in the world. And it requires planning perspective to scenario plan ahead and work backwards, and to allocate compute to the most likely scenarios. And so the path that we're on is incredibly compute intensive.
And so I think we don't know where this is going to shake out. I would just make the point that I think we're going to have some company emerge at some point that has a better algorithmic approach, that's more multi-dimensional than just an LLM. And we're not going to know the true dimensions of this until we see that company emerge.
Well, it's a fascinating topic, one I'm sure we'll be talking about a lot in the next six months to a year. And again, it'll be a topic when we have you back, Matt. Great.
Thank you so much for having me on. Incredibly insightful. So thank you guys for coming in for the great conversation.
You can find more CIO content on our website, ubs.com forward slash views, and on various social media channels. Thanks for joining us from New York City. I'm Edwin Marrero.
Have a great day. Thank you for tuning in. Be sure to visit ubs.com slash studios to view the entire UBS studios suite of podcast channels, along with our video offerings, such as UBS trending.
You can also follow us on Instagram for content highlights at UBS trending. UBS studios is part of the UBS chief investment office within UBS global wealth management. Visit ubs.com slash CIO to view the latest research.
This material has no regard to the specific investment objectives, financial situation, or particular needs of any specific recipient and is published for informational purposes only. For a full legal disclaimer applicable to the independent investment views produced by UBS, please visit our website at ubs.com slash CIO dash disclaimer.