The Zhitong Finance App learned that Bank of America strategist Michael Hartnett believes that as tech giants borrow heavily to build artificial intelligence (AI) infrastructure, US technology corporate bonds continue to be sold off, and yields have climbed to attractive levels, and investors may have ushered in the opportunity to buy on bargain prices. At the same time, this unprecedented AI financing boom is changing the way tech giants allocate capital, and the era of large-scale stock buybacks with abundant cash flow in the past may be coming to an end.
Hartnett pointed out that the cumulative decline in US investment grade technology corporate bonds has been 9% since this year, and the performance compared to the overall investment grade bond index has fallen to its lowest level in 30 years. The main reason why technology bonds continue to be under pressure is that large technology companies are concentrating on pouring into the bond market to build AI infrastructure, leading to a sharp increase in bond supply.
According to the data, tech giants including Alphabet (GOOGL.US), Meta (META.US), Amazon (AMZN.US), Microsoft (MSFT.US), and ORCL.US (ORCL.US) have issued a cumulative total of US$182 billion in bonds since 2026, far higher than the US$17 billion in 2024, and increased more than tenfold in just two years. Such a huge supply of new bonds is putting obvious pressure on the price of technology corporate bonds and driving yields to continue to rise.
The yield on technology bonds soared, and the yield on high-quality corporate bonds rose to 6% to 8%
With AI-related bonds being sold off, the yield on large technology corporate bonds, including the “Big Seven” of US stocks, has risen to a high level. Hartnett pointed out that the yield of some high-credit-rated technology corporate bonds has now reached 6% to 8%, while the yield on AI-related corporate bonds with lower credit ratings has reached double digits.
In the high-yield bond market, AI cloud computing company Coreweave (CRWV.US) bonds maturing in 2031 yielded 14.26%, Galaxy Digital (GLXY.US) bonds due in 2031 yielded 11.44%, and the yield on SoftBank Group bonds maturing 2034 reached 9.58%.
In terms of investment-grade bonds, the yield on Oracle bonds maturing in 2066 reached 8.43%, while Meta bonds maturing in the same year yielded 7.40%. Additionally, Amazon's 2076 bond yield is 6.81%, Broadcom (AVGO.US) bonds due 2056 are 6.75%, Nvidia (NVDA.US) bonds due 2056 are 6.67%, and Alphabet's 2075 bond yield is 6.66%.
Other tech giants' long-term bond yields are also generally at a high level. Among them, Micron Technology (MU.US) bonds due in 2051 yielded 6.27%. AMD (AMD.US) bonds due 2052 have a yield of 6.15%, Microsoft bonds due 2062 have a yield of 6.08%, and Apple (AAPL.US) bonds due 2062 have a yield of 6.06%.
According to Hartnett, the previous sharp decline in technology bonds has made the yield of some high-rated corporate bonds attractive, providing investors with potential allocation opportunities. However, rising bond yields also mean higher financing costs for related companies, and ultra-long-term bonds still face a greater risk of interest rate fluctuations.
AI investment changes the financing model of tech giants, and equity financing reached a total of US$271 billion during the year
In addition to a surge in bond issuance, tech giants are also raising capital through equity financing to support expanding AI capital spending. Bank of America data shows that since this year, while issuing 182 billion US dollars of bonds, large technology companies have also carried out equity financing of 89 billion US dollars, bringing the total scale of bond and equity financing to 271 billion US dollars.
This change marks a major shift in how the tech industry manages its balance sheets.
In the past, large tech companies usually relied on strong operating cash flow to support business expansion and used large amounts of surplus capital for share buybacks. However, as the capital required to build AI data centers, chips, and other infrastructure continues to increase, tech giants are increasingly relying on external financing.
Hartnett specifically mentioned that Meta's recent equity financing arrangement shows that technology companies are beginning to pay more attention to the interests of bond investors and the sustainability of capital structures. He believes that this change may mean that the trend of AI-related bonds outperforming technology stocks for a long time is nearing its end.
The scale of stock buybacks of tech giants plummeted, and the era of large-scale shareholder rewards may come to an end
Notably, the AI investment boom is not only driving rapid growth in the size of tech companies' debt, but is also squeezing their cash to give back to shareholders.
According to Bank of America data, the scale of stock repurchases by large cloud computing and AI infrastructure companies since this year has dropped to 14 billion US dollars at an annualized rate, far lower than the 105 billion US dollars in 2024, a drop of nearly 87%. This means that tech giants are shifting more capital from stock buybacks to AI infrastructure construction.
Over the past few years, continued large-scale share buybacks have been one of the important factors supporting the valuation and earnings per share growth of major US technology stocks. However, against the backdrop of a surge in AI capital expenditure, the amount that companies can use to buy back stocks has declined markedly, which may weaken this factor that has supported the performance of technology stocks for a long time.
At the same time, as demand for bond financing increases, technology companies need to find a new balance between expanding AI investments, maintaining credit ratings, controlling financing costs, and giving back to shareholders.