S&P 500 to Sit 12% Lower by End-2027 as AI Capex Rolls Over, Economists Say

The S&P 500 will trade 12% below its current level by the end of 2027, according to Capital Economics, which predicts AI capital expenditure will slow next year and fall outright in 2028. That would leave the index 17% below its early-October peak.

Chief Economic Adviser Vicky Redwood laid out the case in a note on Friday. AI retains genuine potential to lift productivity, she argued, but the revenue required to sustain current spending rates looks implausible without extraordinary economic disruption.

How large is the spending in question?

A handful of hyperscalers (the largest cloud and data-centre operators) are expected to spend $800 billion this year, double the figure of a year ago. Estimates put their investment at $1 trillion or more in 2027, and many forecasts assume annual capital expenditures will stay above $1 trillion for several years.

Redwood calculated that annual AI-related revenues from external customers would need to grow tenfold by 2030 to justify that spending. The arithmetic allows it. The implications do not: it would require far more severe displacement of existing businesses and labor than anything currently visible.

Her verdict follows. Investment should keep rising next year, but at a slower pace, before falling by around 10% in 2028.

What does the capex turn do to GDP?

Higher AI investment added 0.3 percentage point to US GDP growth last year, Redwood estimated, so a slowdown should trim roughly that much from the recent growth rate. A 10% capex drop might shave a further 0.1% off GDP.

That is a modest haircut by American standards. Taiwan faces real GDP growth around 3 percentage points lower next year, and South Korea around 1 point lower, with additional drag in 2028 for both.

Will markets feel it first?

They will, in Redwood's view. Equity markets should price the looming capex drop well before it lands in the national accounts.

"We expect considerably larger falls in tech-heavy indices," she said. "A tightening of financial conditions driven by factors such as higher energy prices or term premia (the extra yield investors demand for holding long-dated bonds) could also make financing investment hard."

Lower stock prices will also feed through the wealth effect (the link between household portfolio values and consumer spending). The past year's rally lifted GDP by around 0.4%, so a selloff should produce a drag of similar scale.

Why no recession call?

Redwood does not expect an abrupt collapse in AI investment or a severe downturn when the boom ends. Several cushions apply:

• Hyperscalers must keep spending to replace depreciating equipment and exploit economies of scale
• The "winner-takes-most" character of the AI race discourages unilateral retrenchment
• Companies can spread investment declines over several years rather than cutting all at once
• AI adoption should keep expanding, adding about 1% per year to US productivity for several years

"The bottom line for the U.S. economy is that we expect slower GDP growth in 2028, but not a major downturn," Redwood wrote. "Admittedly, recessions have sometimes followed previous capex booms and busts, including in the early 2000s after the dotcom bust. However, a fall in capital spending alone has generally not been sufficient to cause a recession."

Her closing line offers the essential caveat for finance teams drafting 2028 scenarios: "Rather, we expect investment to adjust to a more sustainable pace, with sufficient capacity still being added to support further growth in AI usage."

For CFOs, the message is one of sequencing. Capital budgets may hold near historic highs even as growth rates reverse. Equity markets, less patient than depreciation schedules, will likely mark the turn first.