Wednesday, 21 October 2026 44 calls on file SearchSubscribe
Crypto and finance news. On the record.
Breaking
AI3 min read

BofA AI capital cycle warning cites railroads and fiber

Bank of America strategist Sebastian Raedler says AI spending now exceeds the railroad and dot-com booms as a share of GDP, and past cycles disappointed investors.

BofA AI capital cycle warning cites railroads and fiber
Illustration: Called It

A Bank of America AI capital cycle warning landed just as tech stocks slid on Friday. In a note on Oct. 16, investment strategist Sebastian Raedler argued that AI spending is following the path of earlier booms in which the technology worked but investors lost money, writing that "most major capital cycles of the past have ended in disappointment, even though the underlying technology often lived up to expectations."

What happened

Raedler said 12-month forward consensus capital spending by hyperscalers has climbed to $1 trillion from $330 billion in early 2025. That is about 3% of U.S. GDP, above the 2.2% reached during the 1870s railroad boom and the 1.1% of the 1990s tech boom, according to ADVFN's account of the note.

His concern is monetization. Average token prices, what AI providers charge per unit of text processed, have fallen about 50% since the summer as more competitors entered, including Chinese companies offering open-weight models. Business AI spending data from the Ramp AI Index show spending edging lower since August even as token volumes keep rising.

"This monetization disappointment in the face of intense competition is precisely the failure mode seen in past capital cycles," Raedler wrote, as quoted by 24/7 Wall St. Many railroad builders of the 1870s and fiber companies of the 1990s failed, yet the rails and cables they laid stayed in use for decades.

Why it matters

When unit prices fall faster than volumes grow, revenue stalls even while usage booms. That is the risk for anyone whose sales depend on hyperscaler budgets. Nvidia trades at about 43 times trailing earnings against Microsoft's 29 times, 24/7 Wall St. noted, and Nvidia's fiscal third-quarter revenue guidance of about $108.0 billion depends on those budgets holding. Microsoft is a buyer as well as a seller: its fiscal 2026 capital spending reached $115.95 billion while free cash flow fell to $66.99 billion.

Markets reacted on the day. Alphabet fell 2%, Meta dropped nearly 4% and the Nasdaq 100 lost 2.5%, while the semiconductor index posted its longest losing streak since 2008. Nvidia slid 1.43% to $214.13, and Microsoft fell 1.93% to $529.05, taking its market value below $4 trillion. Even so, Nvidia is still up about 15% this year despite a 6.61% drop in the week to Oct. 16, and Microsoft gained 7.9% over the month. Microsoft's finance chief has said most of its capital spending buys short-lived chips and could slow if demand changes.

The AI capital cycle warning does not forecast a crash. It targets investor returns, and the note stops short of saying the technology will fail. It lands days after Larry Fink's data center warning that halting AI build-outs would tip the U.S. into recession, a reminder of how much of the economy now rides on that spending.

What's next

Earnings will test the thesis quickly. Meta and Tesla report this week, followed by other hyperscalers, and their capital spending guidance will show whether budgets are still rising. Microsoft has guided to about 45% constant-currency Azure growth, and a miss would suggest cheaper tokens are hitting a major seller of AI capacity.

For chip stocks already on a record losing run, covered in our chip index losing streak report, any trim to spending plans would land directly on revenue. Until then, the BofA AI capital cycle warning gives bears a historical framework rather than a timetable.

This article is for information only and is not investment advice.

More from AI

All ai

The Morning Call.

The day's crypto and finance news, one call and one chart. Weekdays at 7am ET.