BofA Says AI Concentration in US Stocks Mirrors Run-Up to 2000 Dot-Com Bust

Bank of America said the concentration in artificial intelligence-related trades in the US stock market resembles the period just before the dot-com bubble burst in 2000.

According to BlockBeats and other outlets on October 4, Michael Hartnett, Bank of America’s chief investment strategist, said in a recent report that the market is geared toward buying AI-related assets such as the Nasdaq 100 and the Magnificent Seven, while shunning gauges with less AI exposure such as the equal-weighted S&P 500. He added that the comparison with 1999 still holds.

In the six months before the dot-com bubble peaked in March 2000, US technology stocks rose more than 40%, while consumer staples fell more than 30%. All other sectors outside technology and telecommunications were also weak. Hartnett said a similar pattern is visible now, with AI and megacap technology stocks leading gains while most stocks face pressure from high interest rates.

Hartnett described AI as "the biggest bubble since railroads." Capital spending by hyperscale cloud companies is projected to climb to 3.5% to 4% of US gross domestic product next year. That would remain below the roughly 5% reached during the 19th-century railroad construction boom.

He said it is too early to conclude that the AI boom has already peaked. Semiconductor prices are still rising, marking a difference from the late stage of the railroad investment boom, when freight rates fell after oversupply emerged. At the same time, railroad investment then benefited from falling Treasury yields, whereas today’s market faces the disadvantage of high interest rates.

Hartnett said investors should gradually raise their bond allocations. He recommended a "buy humiliation" strategy, or buying assets that have been heavily sold off and shunned by the market.

The yield on the US 10-year Treasury recently rose to 5.33%, the highest since 2002. Rising bond yields mean falling bond prices.