The Zhitong Finance App learned that large US technology companies are borrowing on such a large scale that bond investors are beginning to think that some emerging market peers are safer investment choices. The yield on SK Hynix's bonds due in 2031 is currently only 9 basis points higher than Amazon's, only one-third of the same period last year, and hedging the credit risk of this South Korean company is lower than hedging the credit risk of many US tech giants. This shift is not limited to individual companies — risk premiums on the Emerging Market Corporate Debt Index are aligning with similar US benchmarks at an unprecedented rate.
This shows once again that the line between risk in emerging markets and developed markets is becoming increasingly blurred, particularly for wealthier Asian economies such as Taiwan and South Korea. But what is really driving this integration is the surge in AI-related borrowing. US technology companies have borrowed more than 600 billion US dollars globally for capital expenses. It is estimated that by 2030, the scale of financing will reach trillions of dollars, and borrowing costs may continue to rise at that time.
Puneet Sharma, head of market strategy at Zurich Insurance, said: “Interest spreads in emerging markets are likely to narrow further. I wouldn't be surprised if they actually trade through developed markets.”
Sharma said that emerging debt is a “passive beneficiary” of the surge in artificial intelligence lending in the US, and bond valuations of hyperscale companies increasingly reflect expectations that bond issuance will increase dramatically in the next few years.

The most obvious sign of convergence is the narrowing of interest spreads between emerging markets and US corporate bonds and US Treasury bonds. According to J.P. Morgan Chase Index data, the interest rate spread between the two is currently only a few basis points, while the historical average is 30 to 50 basis points.
J.P. Morgan strategists Yang-Myung Hong and Alisa Meyers predict that bond spreads will be consistent by the end of the year. “The supply of investment-grade bonds for emerging market companies is more orderly, mainly from Asia and the Middle East, and local investors continue to support demand,” they said in a report to clients.
Take Tencent and Apple as an example. The risk premium for Tencent's 2031 bonds is roughly the same as Apple's bonds due in the same period, which shows that the market views them similar, even though Moody's rated Apple four levels higher than Tencent.
Similarly, TSMC and Meta Platforms' credit ratings are Aa3, but TSMC's dollar bonds due in 2031 are nearly 20 basis points lower than equivalent bonds on the Meida platform. This is the exact opposite of the situation two years ago, when the trading level of TSMC bonds was about 25 basis points higher than equivalent bonds on the Meida platform.
Nazmeera Moola, chief commercial officer at Ninety One, said such pricing is reasonable because “these companies sell chips, and it's easy to predict their actual production volume.”
Businesses other than emerging tech companies are also beginning to benefit. Suppliers of electricity, metals, and other data center raw materials are also benefiting, and demand for copper alone is expected to double by 2040.

Some investors believe that this trend of interest rate spread convergence is excessive, and point out that higher US Treasury yields may be creating good buying opportunities. Euart MacKerron, a credit analyst at Aegon Asset Management, is one of them; his fund doesn't hold Asian tech stocks because he thinks the spread is too small. “If interest spreads continue to widen, we may consider converting some of our emerging market debt into US bonds,” he said.
Asian AI companies are also likely to face huge financing needs. Chinese internet giants' capital expenditure plans and impending debt are expected to drive bond issuance activities to continue until 2026-2027, with Kuaishou and Tencent entering the US dollar market. South Korea has enabled a number of conglomerates, including Samsung Group and SK Group, to commit to invest $880 billion in the chip and data center sector.
Even so, Dickie Hodges, portfolio manager at Nomura Asset Management, said that the huge number of bonds issued by US hyperscale data centers and uncertainty about future returns on capital expenditure put their bonds at a disadvantage. “Future bond restructuring is likely to increase, and interest spreads for AI borrowers may also widen,” he said.