- Big Tech companies including Alphabet, Meta, Oracle, and Amazon have issued nearly $90 billion in investment-grade bonds in a short span to fund AI data centers and related capex.
- Credit spreads on investment-grade corporate bonds sit near historic lows, leading many large asset managers to cut or short IG exposure because they see risk as underpriced.
- Heavy AI-related borrowing is tying up corporate cash flows and increasing leverage, leaving less buffer if economic conditions worsen or AI returns disappoint.
- Analysts expect up to $1.5 trillion of AI-linked bond issuance over the next five years, which could make AI more than 20% of the IG market and strain investor demand and spreads.
Read More
The surge in debt issuance by Big Tech to fund AI data center expansions has raised alarm among bond investors, particularly for the investment-grade corporate bond market. Major hyperscalers like Alphabet, Meta, Oracle, and Amazon have collectively raised nearly $90 billion in IG debt over just two months (Sept–Oct 2025) for these activities. [2][3] Amazon itself launched a $12 billion bond offering in November 2025—its first in three years—to support AI infrastructure, capex, and debt maturities. [4] Meanwhile, private credit markets are under strain, with fund withdrawal limits being imposed, suggesting emerging liquidity risk. [1][2]
Credit spreads—the extra yield corporations pay over sovereign debt—are a crucial gauge of risk. They are currently hovering just above historical lows in the U.S. (~74 bps) and Europe (~84 bps), raising concerns that they don’t account for elevated risk. [1][3] Some fund managers have responded by reducing exposure to IG credit, shorting it, or entirely exiting certain segments. [1] Others are “taking profits” while spreads are tight, fearing any economic weakening or overhang of supply will force spreads wider. [1]
The scale of AI-financed capex is pushing even cash-rich companies toward debt financing. [3] As internal cash flows become increasingly committed to AI investments—almost reaching or exceeding 90% of post-dividends/share buybacks operating cash flow in some cases—there is less buffer for other corporate obligations. [3] If economic conditions deteriorate, or if AI infrastructure returns fall short of expectations, this leverage could exacerbate pressure on earnings, credit ratings, and funding costs.
Longer term, if AI-focused issuance constitutes as much as 20+% of IG bond supply by 2030, demand from traditional IG investors will face increasing competition from supply, potentially straining absorption capacity. [2] Strategic implications include a shift in investment preferences, increased hedging, pressure on spreads, as well as more scrutiny on capex returns and regulatory oversight of private credit and leveraging practices.
Supporting Notes
- Alphabet, Meta, Oracle, and Amazon issued nearly $90 billion in public IG bonds in Sept–Oct 2025 specifically for AI data centers. [2]
- A ICE-BofA spread index tracking U.S. IG companies is only ~10 bps above 27-year lows (~74 bps), signaling unusually tight credit risk pricing. [1]
- Oracle’s 5-year credit default swap cost rose about 44% in a month, reaching ~87 bps, reflecting rising concern among bond/credit investors. [1]
- Investors managing more than $10 trillion in assets express concern or are actively reducing exposure to IG corporate bonds due to overvaluation and thin risk margins. [1]
- Analysts project AI bond issuance could reach $1.5 trillion over the next five years and might represent over 20% of the IG bond market by 2030 if current patterns continue. [2]
- Capital expenditures for AI in 2025 by Amazon rose 61% YoY in Q3, totaling ~$89.9 billion to date, underscoring sharp ramp-ups in spending and reliance on financing. [4]
Sources
- [1] www.reuters.com (Reuters) — 2025-11-20
- [2] www.reuters.com (Reuters) — 2025-11-24
- [3] www.investing.com (Investing.com) — 2025-11-02
- [4] www.ft.com (Financial Times) — 2025-11-17