How the AI Debt Frenzy Shattered the "Invisible Contract" Between Supercomputing Giants and Investors
Key Points
- Amazon, Meta, and Alphabet (Google's parent company) have all raised their AI capital expenditure guidance in recent earnings reports.
- Supercomputing giants are increasingly using debt instead of cash flow to finance capacity expansion, shifting risk to bondholders.
- Investors are concerned that this speculative expansion faces the dual risks of technological obsolescence and excess supply in the bond market.
Global supercomputing tech giants are significantly ramping up AI capital expenditures and relying more and more on financing from credit markets.
However, investors say this shift is shaking the so-called "fortress balance sheet" status of these large tech companies and tearing up the "implicit contract" they once touted—in the past, speculative AI spending was largely isolated from the debt markets.
Following Amazon, Meta, and Google's parent company Alphabet all sharply raising their full-year capital expenditure plans during earnings season, UBS data shows:
The total capital expenditure of global AI supercomputing giants may exceed $770 billion by 2026, about 23% higher than previously expected.
UBS credit strategists pointed out in a February 18 report that this increase means the supercomputing giants’ borrowing will increase by $40 billion to $50 billion, pushing this year's public bond issuance to $230 billion to $240 billion.
Alastair Catherall, Director of Fixed Income Investments at Mirabaud Asset Management, said that this tilt toward the bond market is radically changing the relationship between supercomputing giants and investors.
"For years, we've been told: these AI expenditures would be funded from their own cash flow—it was an equity risk, speculative in nature, and nothing to worry about from a credit perspective."
Catherall said in an interview with CNBC.
"Now, it seems that implicit contract has changed: we could keep lending to these companies, but AI capital expenditures should still be funded by equity or cash...
Now that capital expenditures have been shifted to the debt market, credit quality issues arise."
"Tipping Point" Reached
Last September, Oracle raised about $18 billion in a single bond market transaction, making it one of the largest bond issuances in history. Other companies quickly followed: Google's parent company Alphabet recently issued about $20 billion in bonds, including a rare 100-year sterling bond.
This has brought the industry's debt burden under more intense scrutiny.
"In the past, everyone treated these AA- or A-rated tech companies as 'cash-plus' level investments. Then all of a sudden, so much debt appears on their balance sheets, and the change is dramatic."
Catherall said, "This shift happened just three or four months ago—all investors are now adapting to a new perspective."
Investors already see trouble ahead. BlackRock says that large tech companies are taking advantage of the current credit issuance "boom" to fill the gap between current investment and future revenues.
"The problem is: increased corporate borrowing further adds to the supply pressure on bond markets that are already struggling to absorb massive public deficits," BlackRock wrote in its weekly market commentary.
The world's largest asset management company pointed out that AI builders mainly tap the US investment-grade market for financing, "so we prefer high-yield and European bonds."
BlackRock also said: "The market's focus has shifted: now the market will ask, how exactly will AI adoption translate into revenue and profit? This process of survival of the fittest is a golden period for active investing."
Over the past six months, as Oracle’s stock price declined, its bond credit default swaps (which provide protection against borrower insolvency) have seen intense volatility.
Catherall also pointed out that Alphabet's planned capital expenditures for next year are close to 50% of its revenue, a level he described as "unheard of."
"Ordinary companies never experience this at any time. We are clearly at a tipping point in the natural cycle."
"Hidden Risks"
Investors are concerned about debt-driven AI overexpansion, mainly because the large data centers at the core of this expansion may become instantly obsolete due to rapid technological iteration (more efficient chips, lower compute demand).
Catherall believes this has far-reaching implications for bondholders.
"If in three years, these Nvidia chips are surpassed by Chinese competitors, and my loan term is five or eight years, but by year three my data center is already obsolete, what then?"
"Is the system accumulating hidden risks? For example, special purpose entities, more asset leasing, more off-balance-sheet operations?
Hidden risks are accumulating, and we don't know if these risks will erupt." Raita Ha told CNBC's "Europe Business Report" on Wednesday.
"But clearly, when investors start discounting future equity market returns, they must remain vigilant."
Editor: Guo Mingyu
Disclaimer: The content of this article solely reflects the author's opinion and does not represent the platform in any capacity. This article is not intended to serve as a reference for making investment decisions.
You may also like
After a week of "bloodbath" in US stocks, valuations now show investment value, but no one dares to say the bottom has arrived.
Last week, U.S. stocks experienced a massive sell-off, but in fact, there have been no fundamental changes in the underlying factors. Multiple valuation indicators now highlight their investment value, and the willingness of contrarian funds to buy the dip should be strong.

The script at the end of 2018: The Federal Reserve continued to raise interest rates, US stocks plunged 20% from the September highs, forcing the Federal Reserve to "pivot" in January.
Goldman Sachs' head of trading believes the current US stock market is repeating the scenario of late 2018, with monetary tightening and equity market pressure. The Goldman Sachs Financial Conditions Index has tightened by about 8 points in two weeks, reaching the highest level in a year. The selloff has broadened from tech stocks to defensive sectors. Gold and equities are declining simultaneously, possibly indicating investors are selling assets to raise cash, making gold a key indicator for observing liquidity stress.

Goldman Sachs Trader “Revisits the 2018 Playbook”: Still Clearly Remembers “On December 24th, Clients Asked Why the Market Dropped So Much—And I Had No Idea”
Brian Garrett pointed out that the current adjustment in the US stock market is similar to the sharp decline in Q4 2018: the S&P 500 has already experienced 42% of the corresponding drop, and the VIX is approaching its peak just before Christmas. If history repeats itself, the bottom of this cycle may appear in January next year. However, the credit market and Federal Reserve policy have yet to provide clear signals of a bottom.
Anthropic (ANT.US) nearly “broke issue price,” but Claude knocks on the doors of “the highest peak in mathematics”! 17 million lines of Lean code used to write a candidate proof for the Riemann Hypothesis
On November 20th, Eastern Time, on its second day of trading, Anthropic's intraday price dropped to a low of $83.05, just 5 cents above the offering price of $83, and finally closed at $88.91, a decline of 8.50%. On the same day, the company's mathematician Levant Alperg announced on his personal website that he and Claude had found a candidate proof for the Riemann hypothesis, which has puzzled the mathematics community for 167 years, and released about 17 million lines of Lean formalized code.

