Bank of America's Hartnett: In the six months leading up to the peak of the dot-com bubble in March 2000, only the technology and communications sectors were rising, a situation strikingly similar to the current one.

Bank of America's Hartnett: In the six months leading up to the peak of the dot-com bubble in March 2000, only the technology and communications sectors were rising, a situation strikingly similar to the current one.

Michael Hartnett, chief investment strategist at Bank of America Securities, warned that the current market structure closely resembles that of the period leading up to the bursting of the dot-com bubble in 2000, and advised investors to start buying bonds on dips.

In his latest Flow Show report, Hartnett points out that in the six months leading up to the dot-com bubble peak in March 2000, the technology sector rose by over 40%, while the consumer staples sector fell by 30%, and all sectors except technology and telecommunications declined. He states that this pattern is "strikingly similar" to the current market trend . Meanwhile, 400 stocks in the S&P 500 have fallen below their 50-day moving averages, and 300 have fallen below their 200-day moving averages, with market gains highly concentrated in artificial intelligence-related sectors, represented by "Mag7".

In the bond market, the 10-year US Treasury yield has risen to 5.33%, a new high since 2002. Hartnett sees this as a buying opportunity, proclaiming the slogan "buy humiliation" and advising clients to start increasing their bond positions . He believes that long-term US Treasury returns have fallen to a century low, and historically, similar extreme low returns have often signaled the arrival of generational buying opportunities.

Market rallies are extremely polarized, and analogous signals from 1999 continue to emerge.

In his report, titled "A Tale of Two Cities," Hartnett describes the current market polarization. He writes that the market is "going long on artificial intelligence" (NDX) while "shorting on AI-irrelevant assets" (the S&P 500 Equal Weighted Index SPW), and that the Mag7, AI, and biotechnology sectors have shown remarkable immunity to the high-interest-rate environment, "making the 1999 analogy still valid."

On the data front, statistics from Goldman Sachs' technology trading desk show that even with the index at historical highs, the median number of individual stocks in the S&P 500 has fallen 16% from its peak, and market breadth has fallen to levels seen during the dot-com bubble. The following day, the percentage of stocks trading above their 200-day moving averages further fell below 50%.

Hartnett cited historical examples showing that in the six months leading up to the March 2000 peak, the technology sector surged by over 40%, while all other sectors declined, with the consumer staples sector falling by as much as 30%. He believes that the current market's sector divergence is almost identical to that period.

AI Capital Expenditure Bubble: A Historical Lesson from the Railway Bubble

Hartnett characterized artificial intelligence as "the biggest bubble since railroads" and provided a detailed retrospective of the bursting of the two railroad bubbles in the 19th century in this week's report.

The first railroad bubble ended in 1873: Between 1861 and 1872, railroad stock prices tripled, the length of U.S. railroads expanded from 35,000 miles to 70,000 miles, and peak capital expenditures reached approximately 5% of GDP. Subsequently, the Franco-Prussian War triggered a credit contraction in the United States, capital flowed back to Europe, and Jay Cooke & Company collapsed in 1873. In the following 12 months, approximately 115 railroad companies went bankrupt.

The second railroad bubble ended in 1881: railroad stocks rose 2.5 times from 1877 to 1881, at one point accounting for 63% of the total market capitalization of the U.S. stock market . Railroad construction quadrupled in four years, resulting in overcapacity, a continuous decline in freight rates, a collapse in revenue and profits, and ultimately triggering a banking crisis and the third longest economic recession in U.S. history (1882-1885).

In contrast to the present, Hartnett points out that the capital expenditure of hyperscale cloud computing companies is expected to reach 3.5% to 4% of GDP in 2027, still lower than the peak level of 5% in the railroad era, and semiconductor prices continue to rise, which is very different from the situation in the railroad era where freight rates fell by 5% every year.

However, he also warned that both railroad bubbles were supported by declining Treasury yields, "which is clearly not the case now."

Furthermore, the end of both railway bubbles was accompanied by credit events, a sharp drop in liquidity, and geopolitically driven capital outflows—all three of which are now reflected in Oracle credit default swaps (CDS), the Australian dollar to Japanese yen exchange rate, and the French bond market.

"Buying a disgrace": Hartnett is bullish on neglected US Treasuries.

Hartnett's stance in the bond market has become increasingly clear.

He acknowledged that a 100 to 200 basis point drop in yields might require a credit event or economic recession as a catalyst, but he believes the current extreme positioning itself constitutes a buying opportunity : asset allocators are generally heavily invested in stocks and shorting bonds, betting on the last wave of gains in US tech stocks and the last wave of increases in US Treasury yields.

Regarding long-term US Treasury bonds, long-duration zero-coupon bonds (ZROZ) have fallen by 65% from their peak in March 2020, and the 10-year rolling yield is -2%, a century low. Hartnett cited historical data, pointing out that the last time long-term returns for stocks and commodities were equally dismal marked the emergence of generational buying opportunities.

He further focused on investment-grade bonds issued by hyperscale cloud computing companies, noting that the U.S. Investment Grade Technology Bond Index (CITE) has fallen 9% in price over the past year, while yields have risen from 4.5% to 6.2%. He believes that given the U.S. government's backing for the AI industry, the long-term bond yields of hyperscale companies such as Oracle (yield 8.4%), Meta (7.5%), and Google (6.9%) are approaching junk bond levels and may soon attract buying interest.

Investors began to increase their bond holdings.

The latest EPFR data (as of Wednesday) shows that bonds saw a net inflow of $18.8 billion, stocks a net inflow of $15.8 billion, cryptocurrencies a net inflow of $900 million, gold a net inflow of $700 million, while cash saw a net outflow of $118 billion due to quarter-end portfolio adjustments.

Specifically, long-term bonds with maturities exceeding six years (government and corporate bonds) saw a net inflow of $7.4 billion in a single week, the largest since May 2025; municipal bonds saw a net inflow of $4.2 billion, a record high since data became available in 2004; European stocks saw a net inflow of $1.3 billion, the largest since February 2026; Chinese stocks saw a net inflow of $3.7 billion, the largest in nearly nine weeks; the technology sector saw a net inflow of $3.3 billion, the largest in nearly five weeks; and utilities saw a net inflow of $1 billion, the largest since December 2025.

It is worth noting that Bank of America's private clients currently hold equity positions as a percentage of assets under management that have risen to a record high of 66.3%, while cash holdings have fallen to a record low of 9.4%, indicating that institutional investors' equity positions are extremely crowded.

Four key warning lines

In his report, Hartnett listed four market risk trigger levels that need to be closely monitored: the Global Financials ETF (IXG) falling below $125, the MOVE bond market volatility index rising above 125, the Mid-Cap ETF (MDY) falling below $666, and the Small-Cap ETF (IJR) falling below $135. He pointed out that once small-cap stocks follow the decline of bank stocks, the cascading risk effect of deleveraging will officially begin, at which point even policy support will be difficult to stop.

The Bank of America Bull & Bear Index fell to 8.8 this week from 9.3 , which Hartnett attributed to widening spreads in high-risk bonds, outflows from high-yield bonds, and a rapid deterioration in global market breadth—the net 27% of global stock indices fell below their 50-day and 200-day moving averages simultaneously, their worst level since March. Despite the decline, the reading of 8.8 remains in "sell" territory .

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.