The AI Economy (Part 3) - What the Dot-Com Bubble Can Teach Investors About AI
In Part 1 of this series, we explored why artificial intelligence has become one of the most transformative technologies of our time and why investors are paying close attention. In Part 2, we looked at what Enron can teach investors about the risks of assigning too much value to future expectations before those expectations become realized earnings. There may be no better historical example of this than the dot-com boom of the late 1990s.
The comparison is not meant to suggest that today's AI companies are the same as the companies that failed during the dot-com crash. There are important differences, mainly that many of today's largest technology companies are established business with realized earnings. The lesson learned is that not every company associated with a major technological advancement would benefit.
The internet changed the world, many of the companies made during the internet boom did not.
When the Internet Changed Everything
The late 1990s were a period of extraordinary technological advancement. The uses for the internet were expanding rapidly, businesses were adopting change, and investors' optimism reflected this. The internet transformed how people communicate, shop, work, consume entertainment, and conduct business. Companies such as Amazon and Google eventually became some of the most valuable businesses in the world.
Investors in the late 1990s did not know which companies would ultimately succeed, but that did not stop them from assigning high values to businesses that were still trying to prove their business models. The Nasdaq rose dramatically during the second half of the 1990s, with internet companies becoming some of the most closely watched investments in the market. From 1996 through 1999, "dot-com" stocks increased more than 200% as investors rushed to participate.
By early 2000, the Nasdaq reached a peak of 5,048.62, then the bubble burst.
Quick Takeaway: The internet really did transform the economy. The mistake was assuming that every company would become a successful investment.
The Problem With "This Time Is Different"
One of the most powerful forces in any market boom is the belief that traditional measures of value no longer apply. During the "dot-com" era, traditional measures such as earnings and profits were increasingly viewed as less important because companies were expected to grow rapidly once they established themselves online.
Some companies had little or no earnings but enormous market values, being valued based on factors like website traffic, potential customers, or the size of the market they hoped to capture. The logic was straightforward, the success of the internet was going to become enormous and any companies positioned in this space would eventually benefit greatly. For many companies, "eventually" never came. Federal Reserve research from the period found that internet IPOs between 1996 and 2000 had high valuations despite relatively low operating income and median sales below $45 million.
A company does not have to be profitable today to become a great investment. Some of today's most successful companies required years of investment before producing substantial profits. The problem occurs when investors pay a price today assuming success tomorrow, and tomorrow never arrives.
Quick Takeaway: A company can have enormous potential and still be a poor investment.
The Bubble Was Bigger Than Dot-Com Companies
It is easy to look back at the dot-com bubble and assume it was simply a collection of internet startups with questionable business models. The excitement surrounding the internet fueled investment across technology, telecommunications, networking, data centers, fiber-optic infrastructure, and other industries that were expected to benefit from increasing internet usage.
Companies invested heavily to build the infrastructure necessary for the anticipated increase in demand. The Federal Reserve has noted that the late-1990s technology boom combined strong economic fundamentals and genuine technological advances with increasingly optimistic expectations about the long-term potential of new technologies. When those expectations were reassessed, the resulting investment bust spread well beyond internet startups.
In the telecommunications industry, equity valuations and capital spending soared during the boom, only to collapse as demand failed to justify the enormous amount of infrastructure that had been built. By the end of 2000, the telecommunications industry's market value had fallen sharply, and the industry was facing significant overcapacity.
This is relevant to the industry surrounding AI today. The AI boom is not simply about companies developing AI software, it involves investments in semiconductors, data centers, networking equipment, electricity generation, cloud computing, and other infrastructure. While we may ultimately require all of that infrastructure, investors need to consider whether the amount being invested today will generate an adequate economic return tomorrow.
Quick Takeaway: Investment bubbles are not always limited to the companies creating the new technology. They can spread to the infrastructure and industries built around expectations of future demand.
When Expectations Change
As earnings forecasts came back to earth, technology stocks began falling sharply. Federal Reserve research from the period described the decline as a reassessment of both the expected earnings of technology and telecommunications companies and the elevated valuations assigned to them.
Companies that were growing saw drops in stock prices, the internet became more entwined in our lives and although the industry itself became enormous, many of its early investors still lost money.
Quick Takeaway: A successful technology does not guarantee successful investments. The technology can continue advancing even while valuations fall.
What This Could Mean for AI
“History Doesn't Repeat Itself, but It Often Rhymes” -Mark Twain
AI may prove to be a significant factor in the economy moving forward. It may change how businesses operate, how employees work, and how consumers interact with technology. Companies are spending heavily on AI infrastructure. Businesses are experimenting with applications. Consumers are using AI tools at an increasing rate. But identifying which specific companies will ultimately be the winners and losers will probably be similar to forecasting which companies would survive the initial dot-com boom.
How much success is already reflected in today's prices? The market is not valuing AI companies based solely on what they earn today. Investors are making assumptions about what these companies may earn years into the future.
In the coming years, these questions will eventually be answered.
- Who will capture the profits?
- How much will customers be willing to pay?
- How much will it cost to provide increasingly powerful AI services?
- How much infrastructure will ultimately be needed?
- Will the economic value created by AI be large enough to justify the valuations investors have already assigned to the companies expected to benefit?
Quick Takeaway: The biggest lesson from the dot-com era is not to dismiss transformative technology. It is to recognize that investors can be right about the technology and still wrong about the price.
The Difference Between the Internet Boom and AI Today
It would be inaccurate to suggest that today's AI market is simply a repeat of 1999. Many of the companies leading today's AI buildout are profitable, established businesses with significant revenues and cash flow. During the dot-com boom, many internet companies had little or no realized earnings and highly speculative revenue prospects. The Federal Reserve has specifically noted this distinction when comparing today's AI-related companies with their dot-com-era counterparts.
At the same time, the existence of profitable companies does not eliminate valuation risk. A profitable company can still be overpriced, a successful company can still disappoint investors and a transformative technology can still lead to overinvestment.
Markets rarely repeat themselves perfectly. We may see similar patterns, and these patterns deserve attention:
- Rapid technological change.
- Extraordinary investor enthusiasm.
- Aggressive capital spending.
- High valuations based on possible future outcomes.
Quick Takeaway: Today's AI market is not the dot-com bubble, but the same investment risks can appear whenever enthusiasm causes expectations to move faster than demonstrated economic results.
What This Means for Investors
As we have discussed, it is possible that artificial intelligence will create enormous economic value over the next decade. It is also possible that some of the companies currently viewed as obvious winners will not ultimately capture as much of that value as investors expect.
For long-term investors, that uncertainty is a reason to focus on portfolio construction rather than trying to identify a single winner. Diversification can help reduce the consequences of being wrong about one company, one technology, or one investment theme. Valuation discipline can help reduce the risk of paying too much for future growth.
It is important to recognize that even when the underlying technology is real, the investment opportunity can still carry significant uncertainty.
Quick Takeaway: Investors do not need to predict whether AI will become the next internet. They need to build portfolios that can participate in its potential while remaining resilient if expectations change.
Key Takeaways
The dot-com era offers one of the clearest examples of the difference between technological progress and investment returns. The internet changed the world, but many of the companies that promised to lead that transformation did not survive. Others survived but took years to justify the expectations investors had placed on them. And a relatively small number ultimately became extraordinarily successful businesses.
The lesson for AI investors is that transformative technologies create both opportunity and uncertainty. Investors should be careful when enthusiasm begins to turn future possibilities into present-day assumptions. A company's potential can be enormous, but the price paid for that potential still matters.
Part 4 Preview
The AI investment story extends well beyond the technology companies developing the models and chips. As AI adoption grows, it could have significant implications for energy demand, data centers, infrastructure, productivity, and the broader economy. In Part 4, we'll look at how the growth of AI could affect these areas and what those changes could mean for businesses and investors.
THE AI ECONOMY (Part 1) | THE AI ECONOMY (Part 2)
Sources
- Federal Reserve Board - How Current Conditions Compare with the Dot-Com Era.
- Federal Reserve Board - What Happened to the New Economy?
- Federal Reserve Board - Initial Public Offerings in Hot and Cold Markets.
- Federal Reserve Bank of New York - Do Stock Price Bubbles Influence Corporate Investment?
- Federal Reserve Bank of St. Louis - Toil and Trouble: Asset Prices and Market Speculation.
- Federal Reserve Bank of Richmond - Boom and Bust in Telecommunications.
- Library of Congress - Business Booms, Busts, & Bubbles: The Dot-Com Bubble.