US Big Tech companies are borrowing heavily to finance their artificial intelligence investments, leading to a significant reordering of risk hierarchies in corporate debt markets. This trend has caused some bond investors to view emerging-market peers as safer bets than certain US tech giants. For instance, SK Hynix Inc. bonds maturing in 2031 now yield just nine basis points more than comparable Amazon.com Inc. bonds, a substantial reduction from the gap observed a year ago. Moreover, it has become cheaper to hedge the credit risk of the South Korean company, SK Hynix, than that of many major US technology firms.
The shift extends beyond individual companies, with risk premiums on emerging-market corporate debt indexes converging with comparable US benchmarks for the first time in history. This convergence signals growing investor concern that the substantial debt issuance related to AI is increasing leverage and credit risk among large US tech companies. Analysts suggest that the relevant signal is not immediate default risk but rather a repricing of capital intensity, as hyperscalers funding AI infrastructure with incremental debt face headwinds from higher interest expenses and potentially reduced free-cash-flow conversion over the next 6-18 months.
Amazon is particularly exposed to these dynamics if its AWS growth does not accelerate sufficiently to absorb the rising depreciation, power, networking, and accelerator commitments. This could lead to a worsening of leverage metrics before reported revenue catches up. In contrast, SK Hynix's relative credit strength is attributed to its position at the memory bottleneck for AI hardware. However, this is not a straightforward secular credit trade, as HBM pricing and utilization are highly sensitive to a small number of AI customers and potential supply responses from competitors like Samsung Electronics and Micron. A memory downcycle or customer inventory correction could reverse SK Hynix's outperformance sharply in the 6-18 month timeframe.
