Investing is by necessity a forward-looking exercise, so it’s natural for investors to fixate on the future, and in particular identifying unfolding trends and opportunities in the world of business. Without question the theme that has captured the attention of global investors in the past year has been the explosive growth of generative AI, and investors have apparently forgiven the tech sector for the painful drawdowns of 2022 and are now pushing the S&P 500 to new highs. But as investors we must determine the degree to which this trend marks a true opportunity in the long-term evolution of technological advances in our economy, and to what extent the rally has been fueled by the sort of irrational exuberance that tends to result in steep losses once economic reality finally brings markets back to earth.
In our view, the most salient point of comparison would be the latter years of the dot-com bubble, where traditional valuation metrics were being roundly dismissed as defunct in an age where the internet threatened to overturn the economic status quo. A famous phrase typically attributed to Mark Twain states that “history doesn’t repeat, but it often rhymes,” so as investors it’s important to determine how the investing setup in 2024 rhymes with what we’ve experienced in the past, but also how it might be truly different.
What rhymes
The broad narrative driving equity indices higher is an easy comparison to make to the late 90’s: an exciting new technology is changing the world and threatens those companies who are too unimaginative to harness it. The potential applications of generative AI are indeed quite remarkable, and there are innumerable ways the technology could be applied to new profit-generating opportunities in the future. This potential has certainly not been lost on investors, and funds have gravitated toward any name with even the most peripheral connection to the technology.
The one name who has the greatest exposure to growth in AI is Nvidia, where the nature of their “inch-deep but mile-wide” GPU architecture is perfectly situated to train AI models. Nvidia’s strategic advantage and impeccable corporate execution in this space is undeniable, but the expectations being placed on the company are astronomical: the nearly $2 trillion valuation being applied to Nvidia implies the company will maintain a near-monopoly profit margin of more than 50% while growing sales immensely. In 2020 Nvidia realized just under $11 billion in revenue for the year; the current price being applied by the market suggests that the company will sell that much every week within the next decade. This type of revenue growth is unprecedented for a multi-billion dollar, and that’s just what’s already in the price! For the stock to rise further would require these expectations to be pushed even higher, and it’s clear that at this point the stock trades primarily off of sentiment and hype rather than fundamentals.
We saw a recent example of this on February 21st, when Nvidia reported earnings that exceeded the already elevated revenue and profit expectations markets had placed on the company. Revenue growth beat expectations by $1.7 billion dollars, bringing in another wave of FOMO-driven investments; overnight the value of the company increased by roughly $275 billion. For context, Coca Cola is a ~$260 billion company, meaning an ~8% revenue beat led to an overnight increase in value that exceeds the size of one of the most storied brands in American history. Nvidia’s market cap now also exceeds Canada’s annual GDP and is only 25% removed from being worth more than every company in the S&P/TSX Composite index combined. We view these figures as a dangerous sign of unfounded and unsustainable exuberance.
And, just like 1999, stock market returns are being driven by a very small number of big winners. The so-called “Magnificent 7” stocks have driven most of the S&P 500’s return, bringing market concentration back to levels not seen since the dot-com bubble. With the gains in those Magnificent 7 stocks being mostly driven by more expensive valuations (as opposed to earnings growth), the result is that the S&P 500 is split in two, with the largest names demanding very rich valuations while the “S&P 493” sits very close to the long-term average Price-to-Earnings ratio.
Figure 1: Gains in the S&P 500 have been so top-heavy that the majority of names in every sector other than tech underperformed the index as a whole last year
Source: Etoro, Financial Times, Gavekal Research
What doesn’t rhyme
Tech and generative AI proponents argue that despite expensive valuations and record levels of market concentration, this time really is different because the growth opportunity provided by AI will be captured by companies that are already very profitable (in contrast to the infamous money losing dot-com names like Pets.com or Webvan). Indeed, these large-cap tech names are immensely profitable, with Apple, Facebook, Google and others pulling in hundreds of billions in revenue annually. They have proven business models with wide competitive moats and consistent revenue growth, and in many respects are now so powerful that they can be considered quasi-sovereign entities. Despite rewarding their employees with some of the highest levels of compensation in the world, they also enjoy enviable profit margins. Big Tech has inextricably entrenched themselves into our daily lives in a way that dot-com companies couldn’t have dreamed of, and it’s irrational to expect that these companies will go bankrupt in the next few years the way hundreds of companies did after the bubble popped in 2000. It’s for this reason that, despite our concerns in this sector, we do not foresee the type of catastrophic collapse of the sector we saw 24 years ago.
Our conclusion
We believe it’s important to make a distinction between the trajectory of growing technological advancements and the investing outcomes that can be derived from those advancements. After all, one can argue that the internet really was just as world-changing as its proponents claimed in the late 90s: it’s an integral part of our daily lives, it’s an increasingly dominant force in the global economy, and many large names who failed to adapt (Blockbuster, Xerox, Kodak, Nokia) were left behind. Nonetheless, tech investments in 1999 and 2000 generally ended in very poor returns, with the NASDAQ realizing a peak to trough decline of 75% in just over a year in 2000/2001. In the same way, we do not see it as incongruent to expect that generative AI will be a truly disruptive technology going forward, but that when we look back a year or two from now, investments made today in this space are unlikely to look attractive in risk-adjusted terms.
It's also worth noting that expecting the largest companies to continue delivering outsized returns has historically been a losing bet – on average, 90% of the 10 largest companies in the S&P 500 have underperformed the broader index 1957 to 2023.
Figure 2: Top-heavy outperformance is an anomaly in the past 7 decades, with the top 10 largest companies in the S&P 500 having underperformed the index materially since 1957
Source: Compustat, Standard & Poor’s, RBC DS Portfolio Advisory Group
So while there are some important distinctions to be made between today’s environment and the euphoric dot-com bubble that formed a quarter century ago, we are noticing too much “rhyming” to be completely comfortable with this rally. This is particularly true when considering the availability of more reasonably priced equities outside of the mega-cap tech names, as well as the excellent tax-efficient returns available in low-risk fixed income investments. For now, we’re happy to take the attractive returns that are available without the whiplash that comes from investing in names where valuations can shift by hundreds of billions of dollars overnight.
-The Chieduch Group
RBC Dominion Securities Inc.* and Royal Bank of Canada are separate corporate entities which are affiliated. *Member-Canadian Investor Protection Fund. RBC Dominion Securities Inc. is a member company of RBC Wealth Management, a business segment of Royal Bank of Canada. ®/™ Trademark(s) of Royal Bank of Canada. Used under license. © 2023 RBC Dominion Securities Inc. All rights reserved.