Are US stocks facing a “gray rhino”? AI bonds are exploding, and tech giants are frantically increasing leverage

The Zhitong Finance App learned that the artificial intelligence boom has driven the global stock market to a record high, but this boom now increasingly relies on complex debt financing, because data centers with huge computing power require huge capital investment, and their financial sustainability has attracted investors' attention, and the market is also beginning to debate whether there is a bubble in the AI boom.

Proponents say that unlike previous market frenzy (such as the internet bubble in the late 1990s), this one is driven by companies that are profitable, well-funded, and have irrefutable business reasons. However, some observers, including the Bank of England, say there are some risk points in the financial system. These risk points are concentrated in the area of assets that are opaque, difficult to trade, and have poor liquidity.

Therefore, debt financing in this AI boom is also critical. However, debt financing is often less transparent than stocks, and can be reflected in intuitive stock price changes. In order to more intuitively observe the current AI bond financing situation, this article summarizes the five major characteristics of the current market.

1) Artificial intelligence investment-grade lending has surged

Big tech companies focused on artificial intelligence issued $75 billion in US investment-grade bonds in September and October alone, more than double the industry's average annual issuance volume of $32 billion between 2015 and 2024, according to Bank of America data. The bond offering included $30 billion from Meta (META.US) and $18 billion from Oracle (ORCL.US). Additionally, Google's parent company Alphabet (GOOGL.US) announced new loans on Monday.

The $75 billion deal reached in September and October accounts for only 5% of the total $1.5 trillion US investment-grade bond issuance so far this year. However, Barclays said that the issuance of AI related technology debt is a key determinant of potential credit market supply in 2026.

Debt also takes on mixed forms. Meta, for example, reached a $27 billion financing agreement with Blue Owl Capital for its largest data center project, which adopted a complex structure so that debt was not part of its own statement.

J.P. Morgan estimates that AI-related companies account for 14% of its investment-grade index, surpassing the US banking industry and becoming the dominant industry.

Mag7's net debt-to-equity ratio has dropped from -22% on October 24 to -16% now. Although net debt is still negative, so total liquidity is still greater than debt, the gap between the two is narrowing, which is not a good sign. It can be seen from this that technology companies are spending capital or increasing debt for capital expenses related to artificial intelligence. The closer this value is to zero (or becomes positive), the more the system will shift from “abundant cash” to “leveraged.”

2) Typical example: Oracle's stock price is soaring, but credit risk is rising

Oracle's stock soared 54% in 2025 and is expected to see the strongest annual increase since 1999. Its artificial intelligence-driven revenue surge has made it one of Wall Street's most valuable companies. However, the company's surge in credit default swaps (CDS) suggests investors are concerned about the US tech giant's level of debt.

Big tech companies that have relied on abundant liquidity for many years are returning to debt financing, and in an environment of high interest rates. Even companies with large amounts of cash are starting to issue bonds at such high interest rates.

3) More junk bonds related to artificial intelligence

Bonds related to artificial intelligence are also beginning to appear in the high-yield bond market, or the “junk bond” market. Such bonds have a higher risk of default, but they also have higher returns.

Last month, Bitcoin miner turned data center operator TeraWulf (WULF.US) issued $3.2 billion in high-yield bonds, rated BB- globally by S&P; while Nvidia-backed AI cloud provider CoreWeave (CRWV.US) issued $2 billion in high-yield bonds in May.

4) The increasing role of private equity credit in artificial intelligence financing

UBS said that rapidly growing private equity credit (provided by entities such as investment firms rather than banks to enterprises) is also increasingly funding artificial intelligence data centers. The bank estimates that AI-related private equity loans could nearly double in 12 months by early 2025.

Such loans provide greater flexibility, but may be more difficult to trade during periods of market turbulence, which may put more pressure on financial markets. Morgan Stanley estimates that by 2028, the private equity market could provide more than half of the $1.5 trillion needed to build a data center.

5) Asset securitization reproduction

Morgan Stanley believes that securitized products such as asset-backed securities (ABS) will also contribute to the growth of the artificial intelligence industry. These products package less liquid assets such as loans, credit card debt, or AI data center rents into tradable securities.

Although digital infrastructure accounts for only 5% (or $80 billion) of the US ABS market (about $1.6 trillion), Bank of America notes that the market has grown more than eight times in less than five years. According to estimates, data centers support 64% of this market. It is expected that by the end of next year, the market size will reach 115 billion US dollars. The main driving force is data center construction.

ABS is a standard financing tool, but people have been wary of it since the 2008 financial crisis, when multi-billion dollar products proved to be backed by non-performing loans and highly illiquid and complex assets, which in turn triggered the subprime mortgage crisis.