The latest report from global brokerage firm Jefferies states that concerns about AI are increasing in the stock market. It said that investors’ doubts about the returns from huge capital expenditure (capex) on AI may increase further in the coming quarters. According to the report, the long-standing leadership role of big tech companies in the US stock market is now showing early signs of weakening.
Nvidia’s share in total market cap decreased
The report said that the share of the four major hyperscaler companies and Nvidia in the total market cap of the S&P 500 has come down to 24.7 percent from the record 27.4 percent as of November 3, 2025 and it is possible to decline further. However, these five companies still account for about 41 percent of the S&P 500’s total gains since the AI theme emerged in early 2023.
How much benefit will the huge investment in AI actually yield?
According to Jefferies, investors are now beginning to question how much the huge investment in AI will actually yield. The report warned that the real financial risk is greater for companies that rely on debt to fund AI capex and data center expansion.
Private loan participation on AI is increasing rapidly
The report said that in the last three years, AI was not called a bubble because most of the investments were coming from companies’ cash, but now the participation of private credit is increasing rapidly. According to a study by the Bank for International Settlements, private debt on AI-related companies already exceeds $200 billion, which could rise to $300 to $600 billion by 2030.
What about data center securitization?
Along with this, concern has also been raised over the rapid increase in securitization in data center financing. It is estimated that the annual issuance of data center securitization could reach $30-40 billion in 2026 and 2027, up from about $27 billion in 2025.
Data center securitization is a financial process in which investable bonds or securities are issued based on future income from the data center (such as rent, service fees or cash flow from long-term contracts).
Who is involved in the AI race?
The report also said that big tech companies like Amazon, Alphabet, Meta and Microsoft could collectively spend up to $650-700 billion in 2026 in the race for AI. A large part of this will go on data centres, chips and AI infrastructure. This unprecedented expenditure has raised concerns among investors about pressure on cash flow, potential negative free cash flow, impact on margins and uncertain returns on investment.
Analysts believe that if the expected earnings from AI investments do not come on time, the risk of overcapacity, valuation pressure and correction in the tech sector may also increase, which may affect the overall direction of the US stock market.
