AI hyperscalers have room for ‘elevated debt issuance’ — even after their recent bond binge, BofA says

Bank of America analysts find the top five AI hyperscalers (Meta, Alphabet, Microsoft, Amazon and Oracle) have ample cash relative to debt, with operating cash flow rising to $577 billion this year from $378 billion in 2023 while debt climbs from $356 billion to $433 billion, driving the debt-to-cash ratio down from 0.94 to 0.75. BofA forecasts operating cash flow to hit $1.1 trillion by 2029 versus capex of $632 billion, supporting continued bond issuance (Meta, Alphabet and Amazon raised tens of billions last month); however Oracle is an outlier with negative free cash flow through 2029, higher FY26 capex (+$15 billion) and rising lease obligations that have pressured its shares. The note cautions that physical supply-chain bottlenecks for data-center equipment (turbines, transformers, cooling, high-voltage gear) could constrain the pace of infrastructure deployment despite abundant capital.

Analysis

Market structure: The five hyperscalers (META, GOOGL, MSFT, AMZN, ORCL) show expanding cash generation (OCF rising from $378B in 2023 to $577B in 2024 and projected $1.1T by 2029) while debt rises modestly (to $433B), lowering the aggregate debt/OCF from 0.94 to 0.75 — which favors equity and IG credit of cash-generators and banks underwriting issuance (e.g., BAC). Oracle is the clear outlier with negative FCF through 2029 and elevated lease obligations, making it the primary loser in the short-to-medium term. Data-center equipment vendors (transformers, turbines, chillers) are winners but capacity-constrained, creating bottlenecks that can temporarily raise capex per MW and extend project timelines by quarters to years.

Risk assessment: Tail risks include rapid rate increases making new issuance costly, regulatory shocks (antitrust or export controls) that cut AI monetization, and operational supply-chain delays that push out capacity for 6–24 months. Immediate risks (days–weeks) are sentiment-driven bond/equity moves after deal headlines; short-term (months) are guidance and capex revisions (Oracle already raised FY26 capex +$15B); long-term (years) are execution of AI monetization and manufacturing cadence. Hidden dependencies: large lease liabilities and third-party data-center partnerships can reprice balance sheets quickly.

Trade implications: Favor concentrated long exposure to MSFT/GOOGL/META (secular AI cashflows) and IG credit of the same, size 2–3% per name, with entry on <5% pullbacks. Implement a pair trade: long MSFT/GOOGL vs short ORCL (short size 1–1.5%) for 3–12 months; hedge with ORCL 12-month put spread. Consider buying 5-year senior notes of META/GOOGL if spread to US Treasury >80–100bp; avoid ORCL credit unless spread >200bp or covenants/lease disclosures improve.

Contrarian angles: The market may underprice near-term pricing power from constrained hardware supply — this could lift hyperscaler margins faster than models assume for 2–4 quarters. Conversely, heavy issuance from cash-rich names is not distress-driven but strategic; buying their high-grade bonds can be superior to equity for yield if spreads compress. Oracle’s sell-off may be overdone if partners (Related Digital) confirm projects; trigger-based re-entry (Oracle guidance improvement or confirmed partner financing) would warrant revisiting a speculative long.

Key Decisions for Investors

• Establish 2–3% long positions each in MSFT, GOOGL and META over the next 4–8 weeks, scaling in on any pullback >3–5% from a 30-day high; use a 10% stop-loss or convert to covered-call income if positions appreciate >20%.
• Execute a relative-value pair: go +2% long (split MSFT/GOOGL) vs −1–1.5% short ORCL, horizon 3–12 months; add to the short if ORCL raises fiscal capex by >$10B or issues incremental weak guidance.
• Allocate 0.5–1% to options hedges: buy 9–12 month call spreads (delta ~0.30) on MSFT/GOOGL to express upside with defined risk, and buy a financed 12-month ORCL put spread (25–40% OTM) to cap cost and capture downside from balance-sheet concerns.
• Deploy 3–5% credit allocation to 5-year senior notes of META/GOOGL if corporate yield > Treasury +80–100bp; avoid ORCL corporate paper unless spread widens >200bp or lease/covenant disclosures materially improve within 90 days.