Stock markets are having a wobbly week, and strategists are pointing to the AI capital expenditure cycle as the main concern. Emmanuel Makonga, an equities strategist at Barclays, said the key question is when the AI capex cycle starts to peak, and that investors are focusing on how much further AI spending can sustain from here, a shift that is causing a "de-rating of the AI trade."
The article argues that this concern may understate how much the AI capex cycle *is* the market. Citigroup strategist David Groman and his team show that next year's consensus forecasts are highly concentrated. A single sector providing more than two-thirds of global earnings growth is unusual. Excluding down years for global EPS, it would be second only to the 2022 energy crisis as the highest level of single-sector earnings concentration in the past 30 years. The top-performing sector typically accounts for about 40 per cent of earnings growth on average.
The concentration is especially striking in the US, where ~78 per cent of expected earnings growth next year comes from the IT sector. The author notes that historical precedents do not closely match current forecasts. Stocks usually do well in years of double-digit earnings growth, but tend to struggle when growth decelerates by 20 percentage points, as is expected here. Higher-than-average concentration is neither good nor bad for broad market performance on its own.
Citi's analysis of years with positive but decelerating EPS growth found that stocks rose two-thirds of the time, though no clear pattern emerged. Whether consensus expectations prove realistic is a separate question. The author concludes that whether 2027 will resemble 2000 or 2006 is anyone's guess.
Source: Financial Times Markets · Summarized by HeadlinesBriefing