Key Takeaways

Amazon, Meta, and Alphabet (Google's parent company) have all raised their AI capital expenditure guidance in recent earnings reports. Supercomputing giants are increasingly financing computing capacity expansion with debt instead of cash flow, transferring risk to bondholders. Investors worry that this speculative expansion faces dual risks: technological obsolescence and a potential oversupply in the bond market.

Global supercomputing technology giants are significantly increasing their AI capital expenditures and becoming more reliant on credit market financing. However, investors indicate this shift is shaking the so-called "fortress balance sheet" status of these large tech companies and tearing up an "unspoken pact" – previously, speculative AI spending was largely insulated from debt markets. Following Amazon, Meta, and Alphabet all substantially raising their full-year capital expenditure plans during earnings season, UBS data shows: Global AI supercomputing giants' total capital expenditure could exceed $770 billion by 2026, approximately 23% higher than previous expectations. UBS credit strategists noted in a February 18th report that this increase implies supercomputing giants' borrowing will grow by $40-$50 billion, pushing public bond issuance this year to $230-$240 billion. Al Ketmorel, Fixed Income Chief Investment Officer at Mirabaud Asset Management, stated that this pivot towards bond markets is dramatically altering the relationship between supercomputing giants and investors. "For years, we've been told: these AI expenditures would be funded by internally generated cash flow – it's equity risk, it's speculative, and from a credit perspective, nothing to worry about," said Ketmorel in an interview with CNBC. "Now, that unspoken pact seems to have changed: we could continue lending to these companies, but AI capex should still be funded by equity or cash... Now, moving capex to the debt market raises questions about credit quality." "The Tipping Point" Has Arrived In September last year, Oracle raised approximately $18 billion in the bond market in one go, becoming one of the largest debt issuances in history. Other companies quickly followed: Alphabet recently issued around $20 billion in bonds, including a rare 100-year sterling bond. This has placed the sector's debt burden under increased scrutiny. "Everyone used to treat these AA- or A-rated tech companies as 'cash-plus' investments. Then suddenly, so much debt appears on the balance sheet; the shift is very dramatic," said Ketmorel. "This pivot only happened about three or four months ago – everyone is adjusting to this new perspective." Investors see trouble ahead. BlackRock stated that large tech companies are utilizing the current credit issuance "boom" to bridge the gap between current investment and future revenue. "The problem is: increased corporate borrowing adds further supply pressure to a bond market already struggling to absorb massive public deficits," BlackRock wrote in its weekly market commentary. The world's largest asset manager pointed out that AI builders are primarily financing through the US investment-grade market, "therefore we prefer high-yield bonds and European debt." BlackRock also said: "The market focus has shifted: the market now questions how exactly AI adoption translates into revenue and profit. This process of separating winners from losers is a golden period for active investment." Over the past six months, as Oracle's stock price declined, its credit default swaps experienced significant volatility. Ketmorel also noted that Alphabet's planned capital expenditure for next year is close to 50% of its revenue, a level he called "unheard of." "This is not something you would see for any normal company at any point in time. We are clearly at a tipping point in the natural cycle." "Hidden Risks" Investor concern about debt-driven AI over-expansion also centers on a core issue: the large data centers central to this expansion could become instantly obsolete due to rapid technological iteration. Ketmorel believes this has profound implications for bondholders. "What if in three years, these Nvidia chips are surpassed by a Chinese competitor, and my loan has a 5 or 8-year term? By year 3, my data center is already obsolete?" "Is there an accumulation of hidden risks within the system? For example, special purpose entities, more asset leasing, more off-balance-sheet operations? Hidden risks are accumulating; we don't know if these risks will materialize," Laith Khalaf stated on Wednesday to CNBC's 'Squawk Box Europe'. "But it's clear that when investors start discounting future equity returns, they must remain vigilant of this."