The trend of investors questioning returns from artificial intelligence (AI) capital expenditure is expected to grow in the coming quarters as the market leadership of Big Tech in the US stock market shows signs of breaking down, according to a report by Jefferies.
The report stated that its base case is that the market leadership of Big Tech in the US stock market is breaking down. It added that the trend of investors starting to question the returns from AI capex has only just started, and there is huge potential for these concerns to grow in the coming quarters.
Jefferies said, “GREED & fear’s base case is that the market leadership of Big Tech in the US stock market is breaking down. GREED & fear’s view is that the trend of investors starting to question the returns from AI capex has only just started. There is huge potential for these concerns to grow in coming quarters.”
The report stated this because the share of the four major hyperscalers and Nvidia as a percentage of the S&P 500’s market capitalisation has declined from a record high of 27.4 per cent on 3 November 2025 to 24.7 per cent.
The report stated that this percentage could fall further. However, these five companies still account for an estimated 41 per cent of the gains in the S&P 500 since the beginning of 2023, when the AI thematic entered the US stock market.
The report noted that while this may be a key issue for the overall American stock market trend, the real financial risks lie in companies that have relied on borrowing to fund AI capex and related data centre expansion.
The report also added that it had refrained from calling AI a bubble in the past three years because most of the capex was funded by cash. However, this is now changing with the growing involvement of private credit in funding AI capex.
There are already more than USD 200 bn of outstanding private credit loans to AI-related companies, which could rise to USD 300-600 bn by 2030, according to a recent study by the Bank for International Settlements.
Jefferies warned that the related surge in securitisation of data centre financing may not have a happy ending. Estimates suggest that annual data centre securitisation issuance could reach USD 30-40 bn in both 2026 and 2027, up from about USD 27bn in 2025.
A major recent concern in AI revolves around the massive capital expenditure plans of Big Tech companies. In 2026, firms such as Amazon, Alphabet (Google), Meta and Microsoft are projected to collectively spend around USD 650-700 billion, mostly on data centres, chips and AI build-outs, in an intense race for dominance.
This unprecedented surge in spending has sparked investor worries about cash flow strain, potential negative free cash flow, margin pressure and uncertain returns on investment, leading to stock sell-offs and fears of overcapacity or an AI bubble reminiscent of past technology hype cycles
