The U.S. tech sector's longest-ever dominance cycle has produced the worst returns in history.

Author: Jim Paulsen

Compiled by Deep潮 TechFlow

DeepChao Summary: The belief that tech stocks always outperform has become almost a creed for a generation of investors, but the author, using a century of data, shows that the current tech-dominated cycle is the longest in history—and yet delivers the poorest returns. Understanding this history can help investors avoid mistaking cyclical advantages for eternal laws.

Tech stocks have led the market for too long, conditioning an entire generation of investors to believe that tech stocks are the quintessential buy-and-hold assets because they “always outperform.” Of course, they experience periodic pullbacks along the way, but these corrections simply present excellent opportunities to buy the dip and add innovative new companies to your overall portfolio.

As shown in the chart below, last Thursday was a classic example of increasing divergence in stock market performance in recent years. In just one day, 10 out of the 11 sectors comprising the S&P 500 declined, with losses ranging from -0.39% to -1.50%, yet the S&P 500 as a whole rose steadily by +0.64%, thanks to a surge of +3.4% in the sole Information Technology sector! Ah, why should us struggling investors buy anything other than tech stocks?

To gain proper perspective on today’s remarkable tech-dominated stock market, it may be insightful to examine the historical cycles of tech leadership and compare how this current cycle stacks up—whether it’s the “best” tech-dominated market in U.S. history or merely average when measured against the past. Has the recent buildup of cloud infrastructure, the mobile and social media revolutions, and the transition into the age of artificial intelligence created the greatest tech-dominated market in U.S. history? Or is the current pace of technological innovation and sustained outperformance of tech stocks simply within normal historical bounds?

Figure 1 shows the relative total return performance of the U.S. technology sector since 1926. Before 1989, the computer, software, and electrical equipment sectors from Kenneth R. French’s publicly available database were used and compared against the U.S. total market index; after 1990, the S&P 500 Information Technology Sector Index was compared against the S&P 500 Total Return Index.

Since 1926, the chart shows six prolonged leadership cycles in which the U.S. technology sector dominated, marked by red dashed lines. I use a qualitative approach to identify these leadership periods, defining them as the span from a low point in relative total return to the next major high point in relative total return, with no significant prolonged underperformance phase occurring在此之前. My goal is not to discuss overall bull or bear markets, but rather to capture sustained periods primarily driven by outperforming technology stocks. Of course, these cycles could be divided differently, but this method effectively highlights the major periods in U.S. stock market history led by innovative companies.

Below is a brief overview of each cycle, mostly derived from fragments of AI-generated text. It makes sense—a history of technology, largely generated by technology!

The first major leadership cycle of the past century occurred between March 1927 and August 1929 (labeled #1 in Figure 1), essentially marking the end of the "Roaring Twenties." During this period, the stock market was driven by transformative technologies and industrial innovations that fundamentally reshaped American life and the economy. Key innovations at the time included: commercial radio and mass communication spearheaded by RCA; mass production of automobiles led by giants like General Motors; aggressive advances in polymers, plastics, and synthetic chemicals driven by DuPont and Union Carbide; and a surge in public excitement for commercial aviation following Charles Lindbergh’s historic transatlantic flight in 1927.

The next leadership cycle lasted from May 1932 to December 1937 (Event #2 in Figure 1), primarily driven by the stabilization of the U.S. banking system following the Great Depression. The United States effectively abandoned the traditional gold standard, devalued the dollar, and injected substantial liquidity into the financial system, halting the severe deflationary spiral that had persisted from 1929 to 1932 and encouraging cheap capital to seek high-growth industrial assets. During this period, the Dow Jones Industrial Average rose nearly 500%, with the widespread adoption of radios serving as a key catalyst—radios became the primary medium for news, entertainment, and political addresses, such as Roosevelt’s “Fireside Chats.” The growing availability of reliable airmail contracts and safer passenger aircraft also advanced aviation and logistics, further fueling technological innovation. Additionally, federal spending under the New Deal on infrastructure, rural electrification, and modernization projects directly benefiting technology and manufacturing, along with the U.S. push toward military modernization, provided further momentum. Finally, increased regulatory transparency and the establishment of the Securities and Exchange Commission (SEC) introduced certified corporate financial disclosures and dismantled insider manipulation rings, helping restore public and institutional trust in Wall Street and further fueling the boom.

The next wave of technological leadership from 1952 to 1960 (labeled #3 in Figure 1) was driven by postwar business innovation, massive Cold War defense spending, and pure speculative fervor. The commercialization of the transistor (marking the transition from bulky vacuum tubes to solid-state electronics), the Cold War and the space race (sparked by the Soviet Union’s launch of Sputnik in 1957, which triggered panic in the U.S.), early computing breakthroughs (the emergence of early mainframe computers to automate complex operations, along with foundational high-level programming languages like FORTRAN and COBOL), and speculative mania fueled by “name game” psychology—similar to the 1990s .com boom or today’s AI hype—where investors became obsessed with company names. Simply adding suffixes like “-tron” or “-tronics” to a company’s name could turn it into a stock market darling.

The tech stock surge from 1964 to 1967 (the fourth event in Figure 1) is often regarded as the pinnacle of Wall Street’s “Go-Go Years.” It was driven by technological innovations of the Space Age, the emergence of aggressive mutual fund trading, and widespread retail speculation. During this brief window, investors completely abandoned dull, old-economy blue-chip stocks in favor of soaring “glamour stocks.” Key drivers of this boom included: the “Space Age” and R&D mania (investors began viewing high R&D spending not as current expenses, but as a “magic bullet” guaranteeing massive future profits—does this sound familiar today, amid the AI capital spending frenzy?); the “electronics craze” (companies saw their stock prices skyrocket simply by adding “-tronics” or “computer” to their names—much like today’s firms labeled as “AI” companies?); the rise of the “gun-slinging” fund managers (before the mid-1960s, mutual funds were managed by conservative, committee-based boards; this era gave birth to a new breed of aggressive individual portfolio managers known as “gun-slingers,” who engaged in massive, rapid trading to chase short-term capital gains rather than long-term dividends. This high-frequency trading injected vast amounts of “hot money” into tech stocks—somewhat analogous to recent meme stock investors, cryptocurrency traders, or gold fever participants?); and the precursors to the “Nifty Fifty” (a small group of elite, highly innovative tech companies became market darlings, leading with enormous gains. Investors treated them as “buy-and-hold forever” stocks—today’s equivalents are FAANG and the Magnificent Seven).

The surge in tech stocks from 1992 to 2000 (the fifth event in Figure 1) ended with the infamous dot-com bubble. It was primarily driven by the commercialization of the internet, unprecedented inflows of venture capital, and widespread retail speculation. Key drivers of this tech boom included: the birth of the World Wide Web, the "grow fast" mentality (where venture capitalists and investors prioritized market share and user growth over traditional financial metrics like profitability or revenue), Y2K fears, the democratization of investing (the rise of online brokerages in the 1990s, coupled with affordable personal computers, enabled individual retail investors to engage in day trading of tech stocks, creating a massive speculative feedback loop), and cheap capital (interest rates remained low for much of this period).

Finally, the contemporary rally in tech stocks from 2006 to 2026 (the sixth in Figure 1) has been driven by four structural shifts: the mobile revolution (2007–2015, as smartphones kept people constantly connected), cloud computing (2006–present, enabling infrastructure-as-a-service and subscription models), digital advertising and platform monopolies (2010s), and finally, artificial intelligence and semiconductors (2023–2026).

Analyzing Figure 1 leads to a crucial conclusion: technology stocks are not a "buy and hold" investment. Given technology stocks' historically significant and prolonged periods of outperformance, it’s easy to assume they should always be held indefinitely. In fact, many investors today may feel this way. Imagine buying IBM or Apple at their initial public offerings and never selling—sounds appealing, but broadly diversified technology stock investments have historically not exhibited "buy and hold" characteristics.

Since 1926, U.S. technology stocks have outperformed the overall market’s total return in 50.5% of all months. This means that buying technology stocks at any given time is essentially like flipping a coin, suggesting that a simple “buy and hold” strategy is unlikely to consistently outperform the broader market. In contrast, correctly timing the bull and bear cycles in technology—i.e., identifying long periods of outperformance and underperformance—could generate significant excess returns. Since 1926, during periods of long-term technology outperformance (highlighted in Figure 1), technology stocks outperformed the broader market in an average of 58.1% of months. In all other months, this figure dropped to 44.5%. Clearly, over the long term, a simple buy-and-hold approach to technology stocks is unlikely to be successful. However, a strategy capable of accurately identifying periods of outperformance and underperformance may yield superior investment results.

As clearly shown in Figure 1, except for the third cycle, each time a leadership cycle in technology stocks ends and they begin to underperform, they continue to underperform until they return roughly to the level at which Figure 1 began in 1926. In other words, despite numerous strong technology-led bull cycles throughout history, technology stocks have repeatedly ended up delivering only market-average returns since their inception. Although they experienced several periods of significant outperformance, by the mid-1930s, technology stocks had again returned to matching only the market’s performance since 1926, and nearly so again in the early 1950s. From 1926 until the early 1990s, technology stocks were, since their inception, “underperformers,” only returning to market-average performance again by 2006. In other words, no matter how impressive a technology stock bull market may be, it is always followed by equally numerous and equally severe bear market cycles. Therefore, a simple “buy and hold” strategy for technology stocks, while potentially delivering impressive medium-term results, has not been successful over the long term. In fact, adopting a long-term buy-and-hold approach to technology stocks likely delivers only market-average returns—with significantly higher volatility.

Second, technology stocks have historically experienced both prolonged periods of outperformance and prolonged periods of underperformance. As shown in Figure 1, the six outperformance cycles averaged 93 months, or 7.75 years. Although significant and painful corrections often occurred during these periods, technology stocks tended to maintain a relatively upward trend over extended durations in these tech bull markets. However, investors should also note that the long-term relative total return underperformance periods for technology stocks have been “longer” and equally severe. Of the five complete long-term underperformance cycles since 1926, the average duration was 132 months, or 11 years!

After such a long and mild period of sustained outperformance by contemporary tech stocks (as shown in Figure 1, this long-term outperformance cycle has persisted since 2006), investors should remember: although tech stocks have historically offered astonishing (isn’t that the word most commonly used today to describe tech stock EPS?) outperformance potential, they have also subjected investors to “biblically bad” and even longer periods of prolonged underperformance. When investing in tech stocks, “buy and hold” (with hope) may lull you into complacency during中期 successes, but to profit over a lifetime, successful “market timing” is essential.

Third, interestingly, technology stocks have performed both well and poorly in nearly any economic environment. There appears to be no consistent relationship between economic conditions and technology stock performance. Technology stocks performed well during the economic boom of the 1920s, the recovery from the Great Depression in the 1930s, and the prosperity of the 1990s. However, they performed poorly during the 1940s despite solid real economic growth; underperformed during the 1960–1965 period when real GDP growth was largely strong; and underperformed during the 1980s despite generally healthy growth. What about inflation? Technology stocks underperformed during the high inflation of World War II, matched the market broadly during the hyperinflation of the 1970s, and performed very well during the inflationary period triggered by the pandemic and the Iran war in 2020.

Unlike the old economy, which was primarily driven by industrial and consumer activity, stocks associated with the new era appear far less constrained by traditional economic indicators—whether today or in the past. These indicators include real growth rates, inflation, unemployment, bond yields, or economic policy. In other words, the technology cycle seems to depend more on the state of innovation. It also depends on the intensity of R&D and other investment spending. It depends on how long and by how much tech stocks outperform or underperform the broader market. It also depends on the degree of investor sentiment toward these stocks—is sentiment overly pessimistic or overly optimistic?

Fourth, whether in a bull or bear market, long-term technology leadership cycles typically begin with a bang! The six cycles marked on the chart—regardless of their starting lows or ending highs—mostly exhibit a “V” shape. This means that both the beginning and end of technology leadership cycles are abrupt, with little to no technical warning. The relative total return bottom suddenly reverses and moves sharply upward, while investors are still pondering whether the previous prolonged tech bear market has truly ended or if the current rally is merely a false start. Similarly, technology leadership cycles often end with a sharp decline. Tops rarely show warning signs like rounded arcs; the downturn typically deteriorates rapidly, while most investors remain excitedly waiting for another “buy the dip” opportunity!

Ultimately, how can investors best “judge” whether tech stocks remain a good long-term buy at any given moment, or are nearing a new long-term decline? This is an extremely difficult decision! My guess is that the best approach involves monitoring a set of indicators. These include: the duration for which tech stocks have served as long-term leaders, and the extent to which they have outperformed the broader market within the current leadership cycle. They also include the level of enthusiasm and aggressiveness displayed in U.S. business behavior—such as investment spending, cash flow usage, debt utilization, and the excitement surrounding future projections. They further encompass the degree to which media attention and market culture have shifted toward the tech boom, reflecting how passionate or short-sighted the media is toward technology. They also include the level of enthusiasm, certainty, and complacency demonstrated by investors. Regardless of the method chosen to make this critical judgment, it has always been—and will likely continue to be—a difficult call. Looking back at Figure 1, it was easy to declare the long-term bull market in tech over in 1935, 1955, 1997, 2020, or 2022. Yet each time, that conclusion proved premature. Similarly, it was equally easy to decide to re-enter tech in 1942, 1948, 1989, or 2002. Each of those decisions turned out to be costly.

How does the current "tech rally" compare to the other five major long-term tech bull markets since 1926?

Compared to past historical rallies, the current technology sector rally stands out in one key aspect: it is the longest technology stock rally in the past 100 years—and by a significant margin. As shown in Figure 2, the average duration of the previous five technology rallies was 63.4 months, or 5.28 years. In fact, the longest rally prior to this one ended in 1960 (the third rally in Figure 1), lasting just 99 months, or 8.25 years. The dot-com bubble rally lasted only 87 months. By contrast, today’s ongoing long-term technology stock rally has now entered its 241st month—slightly over 20 years! This may explain why so many investors wonder whether they need to hold anything beyond technology stocks. It also explains why the technology sector dominates nearly all public investment discussions. After two decades of sustained investment success, for many, the stock market has become synonymous with “technology”—in their minds, hearts, and souls!

As shown in Figure 3, the current long-term tech bull market is also unique in history, as it has generated the smallest excess average annual return relative to the overall stock market. This comparison may surprise many investors. Although technology stocks have proven to be excellent investments over the past several decades, their overall relative performance remains the weakest of all tech leadership cycles over the past 100 years. Since 2006, the current tech rally has generated only a 6% excess average annual market return—less than even the weakest historical tech cycle since 1926.

Perhaps not surprisingly. As shown in Figure 4, given that the overall return stream has been less pronounced than previous long-term tech rallies, the current tech rally has imposed relatively low volatility in total returns on investors. Current relative return volatility is nearly on par with the major tech rallies of the 1950s and 1960s and is far below the relative return volatility seen during the Great Depression cycles or the dot-com bubble cycle. In fact, the dot-com bubble cycle stands alone—it imposed significantly higher relative return volatility on investors than any other cycle since 1926.

Finally, the current market’s unit risk relative to total return is the lowest among all long-term tech rallies since 1926 (Figure 5). The tech rally triggered by IBM’s mainframe introduction in the 1960s, as well as the recovery from the Great Depression in the 1930s, delivered the best unit risk-adjusted returns for investors. The primary reason is that both generated significantly higher overall average annualized total returns.

This piece is more of an observation than a bold judgment on how tech stocks should be handled. Nevertheless, the stock market has been led by the tech sector for two decades. It’s worth taking time to understand the history of tech dominance in the market over the past century. I believe this current cycle has lasted longer than any previous long-term tech rally. One reason is that this cycle has been characterized more by sustained growth than by explosive bursts. Sustained growth may prove to be sustainable, whereas bursts often lead to crashes—at least that has been the pattern until recently, when the AI boom has significantly heightened investor enthusiasm and corporate aggressiveness.

An increasing number of warning signs are making me increasingly uneasy: this cycle is becoming more like a bubble. As shown in Figure 1, the technology sector has just reached a century-high relative to the total return index, suggesting significant downside potential. Its 20-year outperformance is twice that of any previous cycle since 1926. Could technology lead the stock market for an entire 40- to 45-year investment lifetime? Who knows.

The U.S. business community has recently become more enthusiastic about the latest tech craze. Their behavior has also grown more aggressive: burning through cash, accelerating capital expenditure plans, and taking on debt to double down. Financial news streams have almost been hijacked by “tech/innovation stories.” Wall Street analysts now reference data series tracking how often the term “AI” is mentioned. Videos of robots mowing lawns, performing surgery, and even breaking world records in the 100-meter sprint have gone viral. At the very least, after two decades of success, complacency surrounding tech stocks has noticeably intensified—if not become outright extreme.

Tech profits seem to be rocketing toward the moon (or even Mars), and tech stocks have been winning for so long. Who has the guts to truly call a top? I don’t. I might eventually be right, but perhaps only by about five years? Ha! So, while my cautious self isn’t saying the top is near, I will say this: there’s enough cause for concern. Investors should reduce their allocation to the tech/new economy segment of the stock market—not sell everything, but lower exposure.

I’ve left you the chart below, which presents another concerning indicator: the risks in this long-term tech bull market are rising. Figure 6 overlays the U.S. technology sector’s relative total return index since 1955 (blue line, left axis) with the detrended real investment per job ratio (red line, right axis). Bold green numbers mark the peaks of previous long-term tech cycles and the current cycle’s position (originally shown in Figure 1).

Since at least World War II, the relative performance of technology stocks has been closely correlated with companies’ “aggressive” spending per employee. Note that when trend-adjusted real investment per position rises above zero (i.e., investment spending per position exceeds the average), it often signals that a technology-led market cycle is nearing its end. In the third, fourth, and fifth cycles, each cycle ended relatively quickly after investment per position surged above zero. In recent quarters, this ratio has sharply risen above zero, reaching levels comparable to those that preceded the end of the third cycle.

Undoubtedly, the tech rally may continue for some time. However, a series of warning signs are emerging, and the risk of overconcentrating in tech is rising. I’d like to say “be bold.” But with a leading trend that has gone unbroken for two decades still in place, perhaps the most realistic approach is to “be slightly more cautious!”

Thank you for visiting! Jimp

Please note that stocks themselves carry risk. Any stock may lose most of its value at any time, including any stocks mentioned here. You should never rely solely on a single source for investment decisions, including me. I am a retired investment strategist offering insights and observations on markets, the economy, and companies. You should conduct your own research and consult a qualified financial planner or investment advisor before making any investment decisions.

This newsletter is for your reference only. This material is provided for illustrative and discussion purposes only and should not be construed as financial, legal, tax, or investment advice. The information contained herein is current as of the date indicated and reflects only the author’s current expectations. The author may revise this material but is under no obligation to do so. The views expressed herein are solely those of the author. This material should not be interpreted as an offer, solicitation, recommendation, or advice to buy or sell any investment. The information in this report is not intended to serve as a basis for making investment decisions regarding any specific security or its issuer.

Certain information is provided by third parties or based on third-party sources. Although this information is considered reliable, no representations are made regarding its accuracy, completeness, or timeliness. This information may be changed at any time without notice. I assume no obligation to maintain or update this material based on subsequent information or events, nor do I assume any obligation to provide any additional supplementary information, updates, or corrections to the information herein. Any opinions and assumptions expressed in this document are current as of the date of this communication and may be changed or withdrawn at any time without notice. The author assumes no liability for any losses arising from the use of the information contained herein.

Any financial instruments mentioned in this document are inherently speculative and may involve risks to both principal and interest. Any prices or levels shown are historical or purely indicative data. This material does not take into account the specific investment objectives, financial situation, or needs of any particular investor. It is not intended to recommend any specific securities, investment products, or other financial instruments or strategies to any particular client. The securities, investment products, other financial instruments, or strategies discussed herein may not be suitable for all investors. Recipients of this report must make independent decisions regarding any securities, investment products, or other financial instruments.