History often rhymes: the dot-com bubble vs today’s AI euphoria

The Wall Street Journal recently reported that only a small number of portfolio managers active today were also managing money during the dot-com bubble of 1999/2000. I am one of them. At the time, I was the sole manager of the Vontobel US Equity portfolio, and today, I am one of three portfolio managers running the same portfolio. While it has already been written in the press about the similarities and differences between the dot-com bubble then and the potential for an AI bubble today, I would like to share some personal thoughts.

How much certainty does an investor require?

Before diving into a comparison of 1999/2000 and today, let’s take a step back and start with some basics. At Vontobel, we pride ourselves on being investors (as opposed to speculators) who conduct deep research on high-quality companies that we may determine to be "investable," if the price is attractive.

There is a stark difference between an “investment” and a "speculation." A speculation entails buying an asset based on what one thinks the next person will pay for it. Having attended about 30 Berkshire Hathaway meetings in Omaha over the years, I take my definition of what constitutes an investment from my north star, Warren Buffett.

For Buffett, one buys an asset for the cash it will return to the investor over its lifetime and the present value of that is the stock's “intrinsic value.” In his 1993 letter to Berkshire shareholders, he noted that “the certainty with which the long-term economic characteristics of the business can be evaluated" is critical to investing.

So, as we think about the Magnificent 7 companies that constitute almost 40% of the S&P 500 Index today, and the rapidly changing technological environment in which they operate, an important question comes to mind: to what extent can an investor achieve the level of certainty he needs in order to evaluate the long-term economics of these businesses?

With the growing participation of the retail investor in today's stock market, how many are merely speculating versus truly investing? Are they buying these stocks today simply because they believe someone else might pay a higher price tomorrow, or are they doing so simply because of FOMO (fear of missing out)?

The rhyme

In 1999, we held zero percent of the portfolio in tech and telecom, despite those stocks constituting 40% of the S&P 500. At one point, we lagged the market by 30%. Similar to today, there was great pressure to compromise or stretch one's investment approach to "adapt" to the new environment. There was career risk as well, as clients began to sell out of our US portfolio and several well-known portfolio managers were fired…. just before the tech bubble burst.

The reasons for having zero percent exposure to tech in 1999 are similar to why our portfolios are generally underrepresented in AI today. Back then, so many tech and telecom companies massively built up fiber-optic telecom networks but it was by no means clear how strong demand for the networks would ultimately be. Similarly, today, so much datacenter capacity is being constructed without a clear understanding of the extent to which the ultimate users of AI will be able to enhance their productivity.

If the productivity of the end user is not significantly enhanced, the demand for many of these AI investments will be muted and the return on invested capital (ROI) of these AI investments will turn out to be disappointing. While some obvious AI productivity enhancements have manifested themselves in call centers, coding, and elsewhere, a recent McKinsey Global Survey revealed that roughly 80 percent of firms using AI reported that it had no material impact on their profits¹.

Can we know today, with enough certainty to make any AI company "investable," the extent to which AI will ultimately enhance productivity at a given company or the broader economy? The answer is no – it is probably beyond our circle of competence. Could we speculate that AI will be a source of incredible change and guess who the ultimate winners would likely be? We could, but doing so would violate our fundamental discipline as investors.

Other similarities between 2000 and today have already been well documented in the press, such as circular financing deals, exuberant retail investor participation (as mentioned earlier), and the increased use of debt to finance AI investments.

The dissonance

Yet, there are also important differences between the tech/telecom companies of the dot-com era and today’s AI leaders. Many of today’s AI leaders are financing their AI investment using their prodigious cash flows provided by the non-AI portion of their businesses. Many dot-com era companies went public with no profits and very little revenue. Today, AI leaders and the "Big Tech" giants like Microsoft and Alphabet fortunately have existing businesses like ads, cloud, and hardware that can fund their AI investments. 

Valuations were more extreme back then, also. At its peak, the Nasdaq-100 forward P/E ratio was around 60, whereas today, while arguably still expensive, the Nasdaq-100 sits much lower at 26-30x. Lastly, interest rates were increasing in 1999/2000, thereby lowering the ceiling on equity valuations whereas today interest rates are declining. 

How can a disciplined investor adapt to today's environment?

While generally underrepresented in AI, we do have some AI-related investments, which would not be completely immune from a collapse in AI should its ultimate productivity enhancement prove de minimis. We have had "derivative" exposure by investing in such companies as Amphenol and Synopsis. In the Mag 7, we also own Microsoft, Alphabet, Amazon, and Meta, principally because of the strength in their non-AI core businesses. And beyond the spotlight of AI, we have found attractive investments in completely unrelated areas of the market, such as AutoZone (consumer discretionary), Aon (financials), and Union Pacific (industrials).

What does the future hold?

The future, of course, is unknowable, but there are various plausible scenarios. If there is indeed an AI bubble, and if it bursts, we expect that due to our underrepresentation in AI, and the beta of the portfolio well below that of the market, we would likely defend better in a downturn. But if today's AI euphoria persists, it may be challenging for us to fully participate on the upside.

However, we are always open-minded and on the lookout for opportunities. There may be stocks that we previously deemed “uninvestable” because of an uncertain outlook. If we become more comfortable with a given company’s outlook, that company could ultimately become “investable.” If the facts change, we are willing to change our minds.

A third possible scenario is if a ceiling is imposed upon the arguably expensive AI space, leading the market to broaden out into some less expensive, overlooked areas such as auto parts and consumer staples. Such a scenario would likely be a more auspicious one for us. 

The future may be uncertain, but one thing that our clients can be certain of is that, just like 1999/2000, despite the pressure of short-term underperformance and client defections, we will stick to our investment discipline. And, at the heart of that discipline is our commitment to remaining investors (not speculators) and to serving as responsible stewards of our clients’ capital, with a focus on long-term appreciation.