FXT Financial Focus (Asia-Pacific 09/16)BIS Warns of AI Investment Risks

The Bank for International Settlements latest quarterly report shows that market attention around AI is shifting from growth prospects toward investment returns. Major tech companies continue to expand spending on data centers, chips and cloud infrastructure. The BIS previously estimated that AI capital expenditure by the five largest hyperscalers would exceed $1 trillion in 2025–2026. As spending rises, investors are increasingly focused on whether these projects can generate sufficient profits and cash flow.

Changes are already emerging in financing. Large tech companies are issuing more long-term bonds, while credit default swap spreads have widened, indicating that creditors are demanding greater compensation for risk. Meanwhile, issuance of high-yield bonds and leveraged loans has slowed. Capital remains available to financially strong tech giants but has become more cautious toward lower-rated companies.

Private credit is also expanding rapidly into the technology sector. BIS research shows that borrowing by U.S. software and technology companies through private credit surged from about $22 billion in 2010 to more than $1 trillion in 2025, raising its share of total private credit from 22% to 44%. Not all of this financing is AI-related, but AI companies have become an important source of new borrowing demand.

One concern is that some loan spreads have narrowed even as borrower fundamentals have weakened. Tech financing often relies heavily on recurring revenue and intangible assets, leaving repayment capacity vulnerable if AI commercialization falls short of expectations. With interest rates elevated, the cost of building data centers, chips and computing infrastructure is also increasing, putting greater pressure on project returns.

Equity markets have already begun repricing these risks. Since late June, momentum in tech stocks has weakened, with valuation multiples for hyperscalers falling nearly 15% and those for semiconductor manufacturers dropping about 40%. U.S. small caps, traditional sectors and markets outside the U.S. have performed relatively better, suggesting investors are not broadly abandoning risk assets but reducing their concentration in a handful of AI leaders.

Looking ahead, if interest rates remain high for longer, tech companies may place greater emphasis on actual returns from AI projects, while some long-cycle investments with uncertain profitability could slow. Tighter financing conditions may also widen the gap between companies, directing capital toward large firms with stable cash flows while raising funding barriers for smaller AI businesses and potentially accelerating industry consolidation.

From FXT‘s perspective, the BIS is not questioning AI’s long-term potential. The key issue is whether profit growth can keep pace with capital spending and rising debt. If earnings lag behind investment, high valuations and financing pressures could amplify market corrections. If AI applications continue to improve productivity and generate stable cash flow, however, current investment levels may remain manageable, with markets gradually shifting from chasing the AI theme toward identifying companies with genuine earnings potential.

(For more insights into global macroeconomic trends and market developments, please follow FXTs official updates. This information is provided for reference only and does not constitute any form of investment advice.)