Membership is now required to use this feature. To learn more:
View Membership Benefits
Advisor Perspectives welcomes guest contributions. The views presented here do not necessarily represent those of Advisor Perspectives.
Larry Fink, CEO of BlackRock, recently declared: “I do not believe there’s an AI bubble by any [stretch of the] imagination.” We both agree and disagree. We believe AI is a technological game-changer on par with the invention of the computer and the internet, and thus not a bubble. However, it’s quite likely that many AI investments are currently in the midst of a financial bubble.
Our concern is grounded in the work of innovation economist Carlota Perez and her acclaimed book, “Technological Revolutions and Financial Capital: The Dynamics of Bubbles and Golden Ages.” Although written in 2002, well before the AI boom, her book offers a clear explanation of why speculative financial bubbles often accompany significant technological innovations. We have extrapolated her framework to the current AI boom.
Before proceeding, it is worth adding context to Larry Fink’s optimism with a quote from Upton Sinclair: “It is difficult to get a man to understand something when his salary depends on his not understanding it.”
Carlota Perez believes that sustained investor profits aren’t guaranteed by the mere existence of a revolutionary technology that benefits the economy and increases corporate profits. This is especially true when stock prices outpace actual technological benefits, pricing in idealized future outcomes today.
Investors, especially those who invest early in a new technology, often conflate technological progress with successful capital deployment. History shows that substantial capital is often lost in the early stages of investment, even during revolutionary innovations. Examples include:
Perez makes an important distinction between technological capital, which evolves slowly, and financial capital, which is “impatient” and susceptible to faulty narratives.
Technological capital consists of machines, processes, know-how, networks, and skills. It accumulates slowly as innovation is integrated.
Financial capital is impatient and refuses to wait decades for the financial benefits of new technology. Instead, investors price in long-term possibilities as if they are near-term certainties. Perez explains that the speculative phase of a technological revolution is dominated by grand expectations, leverage, and storytelling.
We observe this today in AI-related companies. Investors are pricing stocks as if:
As history proves, none of these assumptions are guaranteed.
A common hallmark of technological shifts is the dismissal of traditional valuation metrics. In the case of AI, investors are told that earnings don’t matter and that metrics like capital expenditures, current margins, or market dominance take precedence.
These arguments echo prior booms:
Consider the following paragraph about the British railway mania in the 1800s from Reuters:
In “Engines that Move Markets: Technology Investing from Railroads to the Internet and Beyond”, Alasdair Nairn writes that tech bubbles are characterised by the emergence of a technology about which extravagant claims can be made with apparent justification. New publications uncritically promote the invention. Entrepreneurs create new companies to meet demand from investors, who suspend normal valuation criteria. The technology is often immature. There follows a huge over-commitment of capital, forcing down potential rates of return.
Today, a company’s involvement with AI has become a justification to ignore fundamentals. We have witnessed stock prices surge solely because executives mention “AI” on earnings calls. For some, massive capital expenditures for AI-related projects are viewed as a guarantee of future success rather than risk.
Today’s circular financing arrangements, as shown in the figure below, are applauded despite their questionable foundation.
Perez would describe this optimistic phase as the period during which financial capital detaches from the technology.
It is worth noting that, unlike the dot-com bubble, some of the largest AI companies — Nvidia, Google, Microsoft, and Meta — have significant revenue and earnings growth. However, many other smaller companies fit Perez’s framework, trading on the promise of future growth.
Another lesson from Perez’s work is that capital expenditures and infrastructure buildouts lag innovation by years or even decades. For example, AI will be used by billions of people worldwide and play a role in almost all businesses; however, the ultimate benefits for most end users will not likely be realized until well after the investments have been made.
The full benefits of computers were not realized until software development caught up. The internet required a series of massive investments before becoming commercially viable. AI is no different — that gap between market expectations and economic reality is where bubbles form.
Speculative bubbles are characterized by market concentration, whereby capital flows disproportionately into a small number of perceived winners. As market breadth narrows, index performance becomes increasingly dependent on a handful of stocks.
This is evident today, as shown in the graphic below. A small group of megacap companies — the Magnificent Seven — dominate AI narratives and capital flows. Investors assume these firms will capture most of the future profits simply because they are early leaders.
However, history suggests that first movers often overinvest and fail, while later entrants benefit from lower costs and more refined business models. Moreover, this future competition spends capital more efficiently, as they have greater insight into what the end technology will look like.
For instance, in 1999, the top search engines included Yahoo, AltaVista, and Lycos. Google didn’t make the top seven. Today, Google handles roughly 90% of global searches, while the former leaders have largely faded into obscurity.
Speculative bubbles pop not because the technology fails, but because capital becomes scarce, profitability disappoints, narratives fail to reflect reality, and, at times, government policies shift.
“The collapse of the bubble marks the turning point that allows the full deployment of the technological revolution to begin,” Carlota Perez wrote in “Technological Revolutions and Financial Capital”.
Importantly, this process does not destroy the technology; it restores discipline to the markets. In turn, this enables capital to flow to the most productive and innovative companies, thereby improving technology.
Perez argues that the most productive phase of a technological revolution occurs after the bubble bursts, when excess capacity exists, costs fall, and widespread adoption begins.
The best long-term returns typically occur after expectations reset, not when enthusiasm peaks.
For instance, Yahoo was the leading search engine in 1999. At the time, the stock was up significantly, clearly pricing in a bright future as the search engine for years to come. Many of those investors lost dearly, as shown below.
Conversely, Google wasn’t on many investors’ radars as its search engine was largely unknown. Patient investors who waited for the dot-com bubble dust to settle could have bought Google at a split-adjusted price of $2.61 in 2004. Currently, it trades above $330, yielding an annualized 26% return over the past 21 years.
The key takeaway from Perez’s work is that investors are inherently impatient. They believe they know with certainty what the future holds, thus justifying lofty valuations. However, no one knows what the future holds — especially not with burgeoning technologies such as AI. This period rewards pundits touting convincing but unproven narratives.
AI is not immune to the speculative cycle Perez lays out in her book. The technology is real, but the investment opportunity is far from understood, and appreciating the stark difference between these two states is essential for risk management. As Perez highlights, the best opportunities in AI may not be in today’s high-fliers but in lesser-known companies — some of which may still be private or not yet formed.
Paying too much for a great idea can yield the same financial result as buying a bad one.
A message from Advisor Perspectives and VettaFi: Discover something new! Click here to register for our upcoming webcasts.
Michael Lebowitz is a portfolio manager with RIA Advisors and author for Real Investment Advice. For more information, contact him at [email protected] or 301.466.1204.
Join RIA Advisors and elevate your career within a deeply experienced team focused on innovation. Our collaborative environment is built on a foundation of advanced technology and effective investment models, designed to enhance your ability to serve clients and grow your practice. Benefit from a supportive culture that encourages professional development and fosters a forward-thinking approach. By joining our team, you’ll be part of a group dedicated to excellence and continuous improvement, empowering you to focus on building meaningful client relationships and pursuing your business ambitions. Discover the advantages of working with our accomplished advisory team by starting your conversation today.
Read more on Advisor Perspectives
