AI Sell-Off Is a Major Risk to the Global Economy, BofA Warns

A lasting AI sell-off is now one of the larger risks to the US and global economy, Bank of America's economists warned in a report on Monday, November 2. Their point is that AI spending and AI stock gains have become a real support for growth, so a sustained correction in the AI trade would hit the economy directly through spending and indirectly through the wealth effect, as stock gains that have propped up consumer spending reverse.

The warning, reported by both Fortune and Yahoo Finance, arrives after a rough three weeks. AI stocks lost about $5 trillion in market value over that stretch and the Nasdaq 100 ended Friday about 9% below its October 6 record, according to Yahoo Finance. It also lands the day before a midterm vote in which data centers are on the ballot in 10 states.

~1.1 pts

AI-linked investment's share of Q3 GDP growth of 3.4% (Yahoo Finance)

48%

Stocks as a share of US household financial assets, a record

$84.95B

Data-center construction, annual rate in August (Census)

41

Local data-center ballot measures on November 3

BofA: the AI trade now props up growth twice

The first channel is investment. Spending on computers, other information-processing equipment and software added a little more than 1 percentage point to third-quarter GDP growth of 3.4%, about a third of the total, according to a Yahoo Finance calculation from Bureau of Economic Analysis data. BofA's economists note that even without AI-related spending the economy grew more than 2% last quarter, up from about 1% at the start of the year, so the rest of the economy is not standing still.

The second channel is wealth. Stocks made up a record 48% of Americans' financial assets in the second quarter, according to Federal Reserve data going back to 1945, compared with 39% at the dot-com peak, Yahoo Finance reported. The richest 10% of households own 88% of stocks, per Federal Reserve data, so a lasting drop in AI stocks would land mostly on the households whose stock gains BofA says have supported spending.

BofA also tied part of the recent rise in interest rates to the borrowing behind the build-out. Its economists said rates are up at least partly because of all the new debt sold to fund AI, a theme we covered this morning with Alphabet's euro bond sale.

Goldman: AI spending is still climbing fast

The same Fortune report pulled together figures from Goldman Sachs that show how large the buildout has become. Goldman's economists estimate AI-related hardware investment is $613 billion above its 2022 level, about 1.9% of GDP. Goldman strategist Ben Snider said the biggest tech companies are on track to spend roughly $800 billion on capex this year, up 96%, with analysts expecting $1.2 trillion next year. Our analysis of AI capex and cash flow breaks down who is paying for it.

Markets have started to separate AI from everything else. Since earnings season began, the S&P 500 has fallen 3% while the S&P 500 excluding AI stocks has risen 5%, Snider said, according to Fortune. Hedge funds are underweight the broader AI trade for the first time since early 2023, Yahoo Finance reported, citing Morgan Stanley prime brokerage data.

Data centers are on the ballot Tuesday

The political side of the boom is also tightening. Voters in 10 states will decide 41 local ballot measures related to data centers on November 3, according to Ballotpedia, as reported by Smart Cities Dive and Fortune. Seventeen of them call for prohibiting data-center development and eight would let local governments permit it. Ohio alone has 21. All six data-center measures already decided this year went against development.

Data-center construction spending hit a preliminary record annual rate of $84.95 billion in August, up 73% from a year earlier, according to Census figures cited by Fortune. Goldman's Alec Phillips expects attention on data centers to fade somewhat after the vote and sees a national moratorium as unlikely under any outcome, Fortune reported. Local votes still decide where the next campuses can be built, and data-center construction is part of the AI investment that BofA says the economy now leans on.

What the AI Bubble Index says

The AI Bubble Index read 62.9 at the October 30 close, in the Normal band, down from 64.1 a week earlier. The Fragility axis, at 68, sits slightly above Euphoria at 66.3, and the systemic risk family reads 68.2. Sentiment and attention remain the hottest family at 91.8, while the demand gap is the coolest at 39.1. The long-run core, which compares today with every day since 1990, stands at 78. Follow each component on the dashboard.

BofA's warning is about how much the economy now depends on AI, which is a different question from whether AI stocks are overvalued. Our guide to AI investment risks covers both, and our report on the $1 trillion tech selloff of October 30 shows how the latest slide began.