Speaker 1 (00:01):
Thank you for your time today. Uh, I'm Kevin Hebner, uh, investment strategist at TD Epoch. And today I'd like to spend some time talking about our latest white paper. Uh, it's titled, is AI Already A Bubble? It's our fourth, uh, in a series on ai. And we'll begin by making the case for why this might be a bubble, and then explain why we think this is different this time and what it means for investors. Now turning to to slide two, uh, we discussed the hype cycle. It appears to be in full throttle leading many to ask if AI is just another bout of irrational exuberance. Uh, my career in the investment business started in 1996, so a few years ago, and since then there have been at least six different bouts of irrational exuberance as we show in the chart on the top left. Um, and so you can look at this, and I think it's obvious to ask, is the Magnificent seven just another bubble waiting to burst?
Speaker 1 (01:04):
And although we don't believe AI's just another over-hyped fad, there's a lot of commentary out there that is reminiscent of previous episodes of Speculative Frenzy. And certainly there's a lot of earnings growth already priced into several of the big tech names, one of the, um, names in the Magnificent Seven. Um, on our numbers, it has to grow earnings by 20% per year over the next 18 years to justify its current valuation that has happened to some companies before, but never for a company that's one of the 10 largest in the world. So there is a real question as to whether today's optimism is justified. On the bottom, uh, left, we have a quote from Sarah Guo. She's, um, founder of conviction, that's an AI focused VC firm, and she also hosts, uh, a wonderful AI podcast, but she mentions that I'm all in on a fundamental bet that AI will drive substantial value going forward, but this is a decade plus transition.
Speaker 1 (02:08):
Investors need to be patient, and that's something we try to make on the chart. On the bottom right, it's typically the case that people generally overestimate the impact of new technologies in the short term, but underestimate their impact over the longer term. Uh, that's always the case with new technologies, but I think it's especially true in the digital economy. Now, turning to the the next slide, we talk a bit about concentration. And this is interesting, um, by any measure, um, the stock market day is very concentrated and that often proceeds a bubble bursting. The chart on the top left looks at the mag magnificent seven over 30% of the s and p 500 market cap. Now, um, the appreciation of the MEG seven is represented 45% of the s and ps gained since 2015 and 66% since the beginning of last year. So pretty enormous impact.
Speaker 1 (03:09):
Um, another way to think about it is the chart on the bottom right. This looks at the top 10 stocks in MSCI us. So it's 29% of market cap of that particular index, and that's the highest it's been since 2000 just before the the tech bubble burst. In fact, we can go back all the way to 1926. I have a chart that goes further and the level is now the highest it's been with only two exceptions. One is the.com bubble in the late nineties, and then the roaring twenties bubble in the 1920s. So the level of concentration today is extreme. So the market cap share is very high, but if you look at earnings, um, the MEG seven only represent 22% of Ford earnings, which is just barely above the, the 30 year mean, which is around 20%. And the gap between those numbers, uh, market share for market cap versus EPS, that's the gray shaded area in the chart on the bottom right.
Speaker 1 (04:06):
That's the biggest. It's been again since the, um, the late nineties tech boom period. And it's about four times the historical norm. So certainly there is reason to be concerned about this. Turning to slide four still with the theme of stretched valuations and crowded positioning. The chart in the top left looks at the Ford PE for the top 10 stocks and MSCI us, the Ford PEs 27. Um, that's way above its historical median of 17 times. And if you look at the remaining 599 companies in M-S-C-I-U-S, they're treating on a multiple of 18, which is a slightly above the historical norm. And then the gap between those two numbers, the top 10 and the bottom 599, that's the grade shaded area that's also quite stretched, four times historical norm at the 92nd percentile. And that's telling us that things are pretty stretched at this point.
Speaker 1 (05:06):
We also know that the most crowded trade in markets today is to be long. The magnificent seven, um, 61% of fund managers are long. This is according to a weekly survey done by the Bank of America now. So that, that sort of makes a case of why one should be worried. What I'll do for the next couple minutes is explain why we think this is quite different from previous bubble periods, the tech bubble in the nineties. Um, it could be, I mentioned in our paper the railroad bubble in the 1840s and so on, but why it's different. So the chart on the right looks at the free cash flow growth for the MEG seven versus the the s and p 500, and you can see a pretty enormous difference over the last decade. The s and p's grown free cash flow by 5% per year, on average, the MEG seven
Speaker 2 (05:57):
By 15%, and then the s and p 4 93. So taking out the seven names is only 3%. So there's a pretty enormous difference. And then the consensus for this year is that even though overall s and p growth, earnings growth is only gonna be, well, it's gonna be a number like 12%, the MEG seven is looking to grow 55%, and that's 11 times the expectation for the rest of the s and p. So there really is, um, earnings growth free cash flow growth beyond behind the exceptional up performance of this group. And broadly, we, we've talked a lot over the last eight years. My co-author, bill Priest, and I are about winner takes most economic superstar firms, and certainly the MEG seven fit that group over the last decade. 15% free cash flow growth on average, 20% operating margins, 22% return on invested capital. Very impressive numbers.
Speaker 2 (06:51):
We're going to some of these in a bit more detail now, but this illustrates the power. When you substitute bits for atoms, you get very powerful capitalized business models and taking advantage of really formidable economies of scale. So turning to the next slide five, we talk a bit about the, well, we've already done free cash flow, uh, operating margins and the return on invested capital. The chart on the top left looks at margins. They've averaged 26% over the last decade. Um, that's, that's a bit better than double, um, the average for the s and p as a whole. And for the s and p 4 93, it's slightly below 10%. So you're getting, um, much higher margins with the MEG seven. And there's certainly no signs that there's a normalization of these margins towards, um, the average for the rest of the s and p, the chart of the bottom right looked at the return invested capital, uh, it's 22%.
Speaker 2 (07:49):
Um, I that's the average over the last decade. It's actually been increasing over the last couple years and it's way above, uh, our estimate of the weighted average cost of capital. And so the difference between the green line and the red line, that that represents value added by this group of companies. Um, so the s and p overall averages, uh, ROIC of seven to 9%, the 4 93, 5 to seven. And you can imagine if your WAC is above 7%, it means a lot of companies are not creating value in this market. And we really would need a wac, sort of similar to what we had last decade during the period of extraordinary loose monetary policy for those companies to be creating value. Um, moreover, uh, as we've shown in some of our research previously, my, my co-author, well my colleague Steve Bleiberg and a couple co-authors wrote a paper, um, the PE ratio users manual in 2019 demonstrating that if you have a higher ROIC, that definitely drives a higher pa e ratio. And really that justifies not just a higher PE
Speaker 3 (08:55):
For tech versus the rest of the market, but also for the US versus Europe and, and other, um, different regions. So I think that's quite important. Uh, moving on to, um, final slide. Uh, show me the money. There's three key takeaways from our paper. One is that an awful lot of earnings growth is currently priced in the tech sector. Secondly, investors can swiftly pivot from euphoria to disillusionment when they think progress, in particular, progress in delivering free cash flow is being too slow. And thirdly, and I think this is the most important one, is we don't think AI is just another overhyped fad. There's a very impressive track record in generating free cash flow margins and, uh, returns. The first two takeaways, I think would be equally applicable to lots of previous periods of, uh, irrational exuberance. The late nineties.com boom, the 1840s <laugh>, that's a long time ago with the railway boom.
Speaker 3 (09:54):
But in history, there have been many, but the third is different. This time we have the MEG seven generating really impressive returns and free cash flow. Um, and our view view is that we think the AI enthusiasm will be rewarded at least for the next year or two. But as the <inaudible> quote emphasized at the beginning, um, of the deck, this is a decade plus transition and investors do need to be patient, uh, beyond the next year or two. We think there's a enormous uncertainty about how AI evolves. So we think it's important that investors are cautious regarding speculative growth or hope stocks that is buying into all the hype. So we favor companies with the demonstrated track record of free cash, high margins and high returns, um, going to investors. And we believe the such companies are most likely to create value and reward investors during this period of extraordinary innovation and disruption.
Speaker 3 (10:53):
Thank you that, that concludes our, our discussion for today. Uh, next month we'll have our fifth piece in the AI series. I'll be focused on the implications of AI for energy infrastructure. The sixth one will be focused on the, the reasons why it's a winner takes most, why economies of scale are so powerful in the digital tech model. Um, please reach out to, to me or my colleagues if you have, uh, any questions or would like to discuss, uh, any of this research, and you can also find it available on our website. Thank you. Have a great day.
