The growing use of debt by major US technology companies to finance artificial intelligence (AI) infrastructure could put upward pressure on global bond yields and raise borrowing costs across ASEAN, according to the ASEAN+3 Macroeconomic Research Office (AMRO).

US Treasury yields serve as a benchmark for borrowing costs in financial markets worldwide. Any sustained increase could therefore spill over into bond markets across ASEAN and the wider ASEAN+3 region, affecting external financing costs and investment flows, AMRO said in a commentary on Tuesday.

The regional macroeconomic surveillance organization said the scale of borrowing by major technology companies, known as hyper scalers, has become an increasingly important factor for financial markets as the AI investment boom gathers pace.

In the first three quarters of 2026, Alphabet, Amazon, Meta, Microsoft and Oracle collectively issued about $160 billion in US dollar-denominated debt, with an issuance-weighted average maturity of 16.5 years.

The amount was equivalent to roughly one-fifth of the US$800 billion raised by the US Treasury through 10-, 20- and 30-year securities over the same period.

The growing supply of long-dated corporate bonds has fueled concerns that hyper scalers are competing with US government securities for investors’ funds, potentially pushing Treasury yields higher.

AMRO’s analysis found evidence supporting this concern. On days when the five companies announced new bond issues, the yield on 10-year US Treasuries rose by an average of about 3.8 basis points more than on other trading days.

The findings were based on an examination of daily US financial market movements between January 2023 and October 2026, focusing on bond deals by the five technology companies.

AMRO analyzed 14 new bond issues, excluding two Microsoft bond exchanges, which involved replacing existing bonds with new ones, and two announcements that coincided with major macroeconomic data releases.

The analysis distinguished between announcement days, when a bond deal was made public, and settlement days, when the bonds were delivered and payment was made.

The results suggest that the impact on Treasury yields was driven mainly by two factors: the additional supply of long-term debt and the broader investment requirements of the AI boom.

AMRO identified three channels through which AI-related investment and financing could influence US Treasury yields.

The first is the duration supply channel. As hyper scalers issue bonds with long maturities, investors must absorb additional duration risk. This could prompt them to demand higher compensation for holding longer-dated securities, raising term premia and, in turn, long-term Treasury yields.

The second is the investment and saving channel. The expansion of AI infrastructure increases demand for capital across the economy. Unless domestic savings or foreign capital inflows rise sufficiently to meet the additional investment needs, equilibrium real interest rates may increase to restore the balance between savings and investment.

The third is the inflation and monetary policy channel. Spending on data centers and related infrastructure could add to near-term demand and price pressures, particularly in areas where electricity supply and construction capacity are constrained.

If investors expect these pressures to lead to tighter monetary policy, expectations of future short-term interest rates could rise, adding to longer-term yields.

However, AMRO’s findings indicate that the first two channels played a more significant role in the market movements observed.

A breakdown of the 10-year Treasury yield showed that real yields increased by about 3.1 basis points more on hyper scaler bond announcement days than on other days, while changes in breakeven inflation were statistically insignificant.

A separate decomposition found that term premia rose by about 2.5 basis points on announcement days, while changes in expected future short-term interest rates were also statistically insignificant.

Taken together, the results suggest that investors demanded greater compensation to absorb the additional supply of long-dated corporate debt, while the increase in real yields reflected market assessments of the investment required to sustain the AI expansion.

By contrast, AMRO found little evidence that inflation expectations or anticipated monetary policy tightening were significant drivers of the observed yield increases.

The organization noted that its estimates could understate the full market response, as investors often anticipate large bond deals and adjust their positions before formal announcements.

This may explain why movements in Treasury yields on settlement days were generally small and statistically insignificant, suggesting that markets respond when companies disclose their financing plans rather than when the bonds are ultimately issued.

AMRO cautioned that the impact of individual deals was modest and cited recent research by the Bank for International Settlements suggesting that the effects could be short-lived.

It also acknowledged that its analysis was based on a relatively small sample of bond issues, meaning further evidence would be needed to assess the longer-term implications.

Nevertheless, the findings suggest that hyper scaler financing should be monitored alongside US Treasury issuance and monetary policy developments, given its potential influence on global financial conditions.

For ASEAN+3 economies, the implications could extend beyond higher government and corporate borrowing costs to changes in portfolio flows and regional bond market conditions. The extent of any spillovers would depend on the financial and economic characteristics of individual economies.

AMRO also warned that risks could emerge in the opposite direction. A sharp slowdown in AI investment or a reversal in investor sentiment could trigger a broader repricing of financial assets and increase market volatility, potentially spilling over into the region.

As AI-related investment continues to expand, policymakers in ASEAN+3 will need to monitor not only the scale of the spending but also how it is financed, particularly as rising corporate borrowing could influence the global benchmark yields that shape the region’s financial conditions.

Asia’s AI boom brings growth opportunities but exposes ASEAN+3 to timing risks, says AMRO