New Delhi

Investor scepticism over returns from artificial intelligence (AI) capital expenditure is expected to intensify in the coming quarters as the long-standing dominance of US Big Tech stocks shows signs of weakening, according to a report by Jefferies.

The report said its base-case assessment suggests that Big Tech’s leadership in the US equity market is beginning to fracture. It added that investors have only recently started questioning the returns on heavy AI-related spending, with significant scope for such concerns to deepen over time.

Jefferies highlighted that the combined market capitalisation share of the four major hyperscalers along with Nvidia within the S&P 500 has declined from a record 27.4 per cent on November 3, 2025, to 24.7 per cent. While this figure could fall further, the five companies still account for an estimated 41 per cent of the index’s gains since early 2023, when the AI investment theme gained momentum.

The report noted that while this shift may influence broader market trends, the more significant financial risks are emerging among companies that have relied on borrowing to fund AI-related capital expenditure and data centre expansion.

Jefferies said it has refrained from labelling AI as a bubble over the past three years, largely because most investment was funded through internal cash flows. However, this dynamic is changing as private credit plays an increasing role in financing AI infrastructure.

Citing a study by the Bank for International Settlements, the report noted that outstanding private credit loans to AI-linked firms already exceed USD 200 billion and could rise to between USD 300 billion and USD 600 billion by 2030.

The investment bank also cautioned that the growing securitisation of data centre financing could pose risks, warning that annual issuance could climb to USD 30–40 billion in both 2026 and 2027, compared to about USD 27 billion in 2025.

Concerns have been amplified by the scale of planned spending by leading technology firms. In 2026, companies such as Amazon, Alphabet, Meta and Microsoft are expected to collectively invest USD 650–700 billion, largely on data centres, chips and AI infrastructure.

READ MORE: When Zafar Sareshwallah offered namaz in RSS headquarters

The unprecedented scale of this spending has heightened investor concerns around cash flow pressures, potential negative free cash flow, margin compression and uncertain returns, fuelling fears of overcapacity and a possible AI bubble similar to previous technology-driven boom-and-bust cycles.