"When the Bubble Bursts, the Bill Comes Due": BIS Sounds Alarm Over AI Euphoria as Big Tech Borrowing Race Raises Financial Market Risks

BIS Warns AI Boom Could Tighten Financial Credit Conditions
Growing Reliance on Private Credit and Bond Markets by Big Tech Emerges as Key Risk
Alphabet Raises Massive Capital Through Bond Markets, Expands Funding With Private Equity Partnerships

The Bank for International Settlements (BIS) has issued a warning over the artificial intelligence (AI) investment boom. While optimism surrounding generative AI has fueled massive capital expenditures, the institution cautioned that if commercialization fails to keep pace with market expectations, a sharp tightening of credit conditions could follow. Market participants are also increasingly warning that the rapid rise in AI infrastructure financing through private credit and corporate bond markets is amplifying risks across the broader financial system.

BIS Examines the AI Investment Boom

According to a Financial Times report published on June 28 (local time), the BIS stated in its latest Annual Economic Report that "today's AI optimism risks being followed by a prolonged investment slump." The institution warned that if returns on AI investments fall short of market expectations, investors could sharply scale back funding, leading to deteriorating credit conditions throughout financial markets. Global AI spending has recently accelerated at an unprecedented pace. Microsoft, Amazon, Alphabet, Meta, and Oracle are collectively expected to spend more than $1 trillion on AI infrastructure between 2025 and 2026.

The BIS argued that today's AI investment frenzy bears striking similarities to past speculative booms that reshaped the global economy, including the canal construction boom of the 1830s, Britain's railway investment surge of the 1840s, and the late-1990s dot-com bubble. In each case, transformative technologies attracted enormous capital inflows, but excessive investment beyond underlying profitability ultimately triggered collapsing investment and economic downturns. The BIS warned that the fallout from an AI bubble could prove even more severe than previous episodes, citing the significantly larger share of equities in household wealth today and the increased vulnerability of the financial system resulting from large-scale debt issuance by AI companies.

The BIS also identified growing dependence on private credit provided by non-bank financial institutions as a major source of concern. According to the BIS, private credit lending to AI companies surged from $3 billion in 2010 to $40 billion last year. Because private credit markets involve a complex web of investment funds, insurers, pension funds, and asset managers, while loan terms and asset valuations remain largely opaque outside public markets, it is difficult to determine where and to what extent potential losses could spread. Unlike traditional banks, these institutions lack deposit-based funding and central bank liquidity backstops, leaving forced asset sales as one of the few options available should investors demand large-scale redemptions. The BIS warned that if enthusiasm surrounding AI fades and investors belatedly reassess risks, declining asset values combined with funding withdrawals could transmit financial stress across the broader market.

Bond Market Emerges as Another Source of Financial Risk

The corporate bond market has likewise emerged as another area of concern. Traditionally, major technology companies financed new business investments primarily through robust operating cash flows. However, as AI spending has begun to outpace cash flow growth, the financing landscape has shifted. In an effort to avoid falling behind in the race to build AI infrastructure, major technology firms have aggressively increased bond issuance, making public credit markets a critical funding source. Citing data from S&P Global Market Intelligence, Axios reported that Alphabet, Amazon, Meta, Microsoft, and Oracle collectively raised $255.34 billion through debt and equity offerings during the first half of this year alone.

Unlike private credit, bonds trade in public markets, allowing for faster price discovery and greater transparency. However, this also means shifts in investor sentiment are immediately reflected in yields and bond prices. Should AI investments fail to deliver expected returns or data center utilization rates disappoint, investors are likely to demand higher yields while outstanding bond prices decline. Any deterioration in corporate credit ratings would further increase refinancing costs. One market expert noted, "Data centers are assets subject to rapid depreciation and are highly sensitive to power availability, cooling capacity, and semiconductor supply bottlenecks. The longer it takes to recover investments, the heavier the borrowing burden in the bond market becomes."

Problems could also emerge if hyperscalers slow or halt capital expenditures. Companies that expanded debt on expectations of continued AI investment—including data center contractors, power equipment suppliers, and server and semiconductor manufacturers—would likely face weakening revenues alongside mounting repayment obligations. The BIS noted that "credit spreads for some companies within the AI supply chain have already begun widening," adding that this contrasts sharply with equity markets, which continue to price in substantial upside potential. According to the report, credit default swap (CDS) spreads for AI-related companies rated BBB or higher have widened noticeably since the first quarter of this year, briefly approaching an increase of 20 basis points in March. CDS contracts provide protection against corporate default risk, and widening CDS spreads generally indicate that investors are becoming increasingly concerned about a company's ability to meet its debt obligations.

Alphabet Accelerates External Fundraising

Alphabet stands out as one of the most prominent examples of companies pursuing aggressive external financing strategies. In April, Alphabet raised its projected annual capital expenditures for this year to between $180 billion and $190 billion, significantly above the $175 billion to $185 billion range announced in February. The company said most of the additional spending would be directed toward technology infrastructure, including servers and data centers.

The key issue is that Alphabet is financing much of this investment not through operating profits but through external fundraising. Its primary funding source has been the bond market. In November last year, Alphabet raised $17.5 billion in the United States and an additional $7.44 billion through European bond offerings. In February, the company launched another bond sale initially targeting $15 billion. Investor demand exceeded $100 billion, prompting Alphabet to increase the final offering size to $20 billion while securing lower-than-expected borrowing costs. The company subsequently issued sterling- and Swiss franc-denominated bonds. Its $1.32 billion equivalent 100-year sterling bond attracted nearly $12.54 billion in investor orders.

Alphabet also disclosed on June 5 that it had raised approximately $17 billion through bond offerings denominated in euros and Canadian dollars. During the same month, Reuters and Nikkei reported that Alphabet was preparing to issue yen-denominated corporate bonds, commonly known as Samurai bonds. The transaction is being arranged by Bank of America, Mizuho Securities' U.S. subsidiary, and Morgan Stanley. Although the final issuance size has yet to be disclosed, market participants expect it to reach several billion dollars, with pricing details expected to be finalized in the near future.

In addition, Google has established a financing structure that shares part of its AI infrastructure investment burden with external investors. The company is launching a major joint venture with global private equity firm Blackstone. Under the plan announced last month, Blackstone will make an initial equity investment of $5 billion and ultimately support approximately $25 billion in computing infrastructure investment, including debt financing. Google will contribute its Tensor Processing Unit (TPU) hardware, software, and cloud services. Through the partnership, the two companies aim to secure 500 megawatts (MW) of data center power capacity by 2027.