Goldman Sachs & Barclays: U.S. Stock Earnings Growth to Be Halved by 2027
nashnova research
Multiple Wall Street firms are warning in unison: global earnings growth will roughly halve in 2027, as AI capex growth peaks — yet slowing profits don't necessarily mean a market top, with valuations already pulled back.
How steep is the earnings slowdown?
Bloomberg consensus puts MSCI World 2027 earnings growth at 16%, nearly half the 33.5% expected for 2026.
Goldman forecasts US earnings growth dropping from 36% in 2026 to 11% in 2027; Barclays sees a decline from over 30% to 15%.
Asia-Pacific ex-Japan faces the sharpest swing: from 74% — the highest since the dot-com bubble — down to 26%, driven by the heavy index weight of chip giants like SK Hynix and TSMC.
Why is growth falling off a cliff?
The core reason: AI capex growth is peaking. Goldman expects global AI investment to rise from $1.1 trillion in 2026 to $1.8 trillion in 2027 — still growing in absolute terms, but the rate itself is slowing.
This means → hyperscalers (the mega cloud companies like Microsoft and Google) are spending on AI infrastructure at a decelerating pace, which directly drags on upstream chipmakers' and memory firms' profit growth.
Two hidden bills are also coming due: depreciation on physical AI assets (servers, data centers) is eating into margins, and borrowing costs have surged after the Middle East conflict drove up energy prices and inflation — expenses that will hit income statements more fully in 2027.
What does Barclays say about the AI trade?
Barclays US equity strategist Venu Krishna put it bluntly: the AI trade — the core engine of US equity outperformance over the past four years — faces its toughest test yet.
He noted that hyperscaler financing has become "more complex, more leveraged, and more market-sensitive."
In plain terms = these companies borrowed easily and spent freely for years; now financing conditions have tightened, and spending momentum naturally slows with them.
Is Europe slowing for the same reason?
No. Europe's earnings deceleration stems mainly from energy-sector pullback, not an AI peak.
Surging oil and gas prices this year created an unusually high earnings base; Goldman expects Brent crude to fall to roughly $78 per barrel by end-2027, down from above $90 currently.
Goldman sees Stoxx 600 earnings growth dropping from 18% in 2026 to 7% in 2027; Morgan Stanley forecasts a decline from about 22% to 7.5%.
Will slower earnings tank the market?
Most firms say no. Morgan Stanley expects both the Stoxx 600 and the S&P 500 to rise 15% in 2027; Barclays targets the S&P 500 at 8,700 by end-2027, up roughly 9% from the projected 2026 year-end level of 7,950. The index currently sits around 7,325.
Goldman's Oppenheimer argues that despite slower earnings growth, US equities are "relatively attractive again" because "valuations have pulled back sharply." Goldman recommends overweight positions in US and Japanese equities for 2027.
This means → firms believe the "too expensive" problem of recent years has been partially digested — even if profit growth slows, prices don't necessarily follow it down.
What picks up the baton next?
The key validation point: once AI capex growth peaks, can AI's productivity gains spread from infrastructure builders to the broader universe of adopters?
In plain terms = the past few years' money went into "building the road" — data centers, chips. The next phase depends on "cars running on that road" — businesses actually using AI to boost efficiency and generate profits, to sustain the next leg of earnings growth.
This reflects a shifting Wall Street consensus: from "who is building AI" to "who can make money with AI."
市场有风险,内容仅供研究参考,不构成投资建议。
