AI is aggressively "leveraging up"

```

The grand narrative of the artificial intelligence revolution is being built on an ever-growing pile of debt. To meet AI’s astonishing demand for computing power, a construction boom in data centers is being fueled by record-scale debt financing, with increasingly complex financial structures that are prompting investors and regulators to be wary of potential bubble risks.

According to Reuters, the latest news shows that in just September and October of this year, technology giants focused on AI issued $75 billion worth of bonds in the U.S. investment-grade bond market, injecting massive capital into the industry. This scale is more than double the industry's annual average bond issuance between 2015 and 2024.

Amid this frenzy, financing channels have extended beyond traditional investment-grade bonds to include riskier high-yield bonds, less transparent private lending, and structured financial products that played controversial roles in the financial crisis. Some observers, including the Bank of England, have already raised warnings that risks are accumulating in the corners of the financial system comprised of opaque, illiquid assets that are difficult to trade.

Though supporters emphasize that, unlike the internet bubble of the late 1990s, this boom is led by profitable, cash-rich companies with clear commercial prospects. Still, from the divergence between Oracle's stock price and credit default swaps (CDS), to Meta's use of complex off-balance-sheet structures to finance data centers, a series of signals indicate that the market is closely scrutinizing the level of leverage behind the AI feast.

Tech Giants Lead Debt Issuance Frenzy, Market Signals Diverge

To fund the AI race, tech giants are flooding into debt markets at an unprecedented rate. According to Bank of America data, in September and October this year, the issuance of AI-related U.S. investment-grade bonds reached $75 billion, far exceeding the $32 billion annual average from 2015 to 2024. This includes $30 billion from Meta and $18 billion from Oracle. Additionally, Alphabet—the parent company of Google—announced a new borrowing plan this Monday, and Bloomberg reports that Oracle’s Vantage Data Centers have secured a $38 billion senior loan.

While the $75 billion issued only accounts for 5% of the $1.5 trillion total U.S. investment-grade bond issuance so far this year, its growth is remarkable. Barclays points out that AI-related tech debt issuance has become a key determinant in forecasting total credit market supply for 2026. Meanwhile, debt structures are growing increasingly complex: for example, Meta reached a $27 billion financing deal with Blue Owl Capital, utilizing a complicated off-balance-sheet structure. J.P. Morgan estimates that in the investment-grade indices it tracks, AI-related companies now make up 14% of the weighting, surpassing U.S. banking and becoming the dominant sector.

Driven by a surge in AI-generated revenue, Oracle’s stock soared 54% in 2025, poised to achieve its strongest annual gain since 1999, making it one of Wall Street’s most valuable companies. However, the cost of credit default swaps (CDS)—which measure the default risk of its bonds—increased over the same period. The rising price of these insurance products indicates that, despite stock market exuberance, bond investors are increasingly concerned about Oracle’s mounting debt levels. The divergence between stock price and credit risk indicators is a microcosm reflecting market anxiety about the “leveraging up” phenomenon in AI.

Financing Channels Expand to “Junk Bonds” and Private Credit

The financing needs of AI data centers are extending capital into riskier areas. The high-yield (“junk bond”) market is beginning to see AI-related issuances. Last month, TeraWulf, a Bitcoin miner turned data center operator, issued $3.2 billion in high-yield bonds, receiving a BB- rating from S&P Global. In May, AI cloud service provider CoreWeave—backed by Nvidia—also issued $2 billion in high-yield bonds.

At the same time, the rapidly growing private credit market has become a key source of funds. According to UBS estimates, in the 12 months leading up to early 2025, the scale of AI-related private credit loans may have nearly doubled. While these loans offer greater flexibility, they can be much harder to trade in times of market turmoil, increasing systemic financial pressure. Morgan Stanley predicts that by 2028, the private credit market could provide more than half the funding needed for the world’s $1.5 trillion data center construction, representing an $800 billion investment opportunity.

Structured Products: “Old Wine in New Bottles”?

Morgan Stanley believes that structured products like asset-backed securities (ABS) will also drive growth in the AI industry. These products package loans, credit card debts, and other illiquid assets into tradable securities. In the AI sector, this means securitizing future cash flows, such as rents paid by tech giants to data center owners.

Bank of America notes that while digital infrastructure currently accounts for only about 5% (roughly $80 billion) of the U.S.'s $1.6 trillion ABS market, its scale has expanded more than eightfold in just under five years. The bank estimates that data centers support 64% of those transactions, and forecasts that by the end of next year, this market will reach $115 billion on the back of data center construction. Though ABS itself is a standard financing tool, it remains under close scrutiny for its role in the 2008 financial crisis, when many such products became worthless due to deterioration in underlying assets, triggering systemic crisis. Today, it is being repurposed to power a new technological revolution.

Risk Warning and DisclaimerThe market has risks and investment requires caution. This article does not constitute personal investment advice and does not consider the specific investment objectives, financial situation, or needs of any individual user. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. You are responsible for any investment made on this basis. ```