Capital Economics: AI Profits May Undershoot Expectations, Weighing on Capital Expenditure
N.R. Finch
Capital Economics warns that AI-related earnings may start missing expectations, with the S&P 500 projected to fall to 6,500 by end-2027 — a signal that the valuation logic underpinning the AI supply chain faces systemic pressure.
Why might AI monetize slower than expected?
Capital Economics sees AI demand still growing, but at a pace below market expectations.
The core reason: companies have broadly underestimated the real-world barriers to deploying AI at scale. This means → the gap between "the tech works" and "the business makes money" is wider than most investors assume.
At the same time, rising competition and rapid innovation are pushing AI service prices down, squeezing revenue growth. In plain terms = the product keeps getting easier to build, but also harder to charge for.
Can stocks still climb — and where does it end?
The research note sees limited further upside in the near term.
But a pullback is coming: the S&P 500 is projected to hit 6,500 by the end of 2027.
This means → the current rally is driven more by momentum than fundamentals, and the turning point may be closer than the market perceives.
Where will the hyperscalers spend next?
Capital Economics expects hyperscalers — Amazon, Microsoft, Google and peers — to prioritize projects with clearer near-term payoffs, rather than continuing aggressive bets on long-horizon AI infrastructure.
This means → AI capex expansion may slow — not a halt, but a shift toward pickier spending.
This reflects a deeper risk: the core thesis supporting AI supply-chain valuations — "the giants will keep pouring money in" — is being reassessed. In plain terms = if the buyers start hesitating, the pricing foundation for the entire chain loosens.
Content is for reference only, not financial advice.