AI-related bond supply has not yet provided evidence of “crowding out” in credit markets
AI summary card
AI-related bond supply has not yet provided evidence of “crowding out” in credit markets
Goldman Sachs believes that the AI financing boom has increased long-duration corporate bond supply, but its crowding-out effect on US Treasuries and non-AI corporate financing remains limited; the more relevant relative value comparison is likely between IG corporate bonds and high-quality structured products.
- AI-related supply has raised IG corporate bonds’ share of DV01 issuance relative to US Treasuries, but this ratio has been higher in the past, and fiscal and macro factors remain the dominant drivers of Treasury yields.
- There is strong investor segmentation between corporate bonds and US Treasuries, making a shift from corporate bonds into Treasuries due to supply technicals less likely than market narratives suggest.
- If investors reduce allocations to IG corporate bonds, the more natural substitutes may be high-quality structured products such as IG CLOs, Agency MBS, or non-QM RMBS, rather than a direct shift into US Treasuries.
- Non-AI corporate issuance remains strong, bank issuance is running at a record pace, and new issue concessions for non-AI deals remain stable, indicating that AI financing has not yet clearly crowded out financing for other corporate sectors.
- The report expects only modest spread widening, as demand driven by high yields and a relatively benign macro backdrop should limit the extent of index-level widening.
Report interpretation
Overview
This report examines whether the AI infrastructure financing boom is creating a new “crowding out” effect in credit markets. Goldman Sachs discusses four questions: whether AI-related debt financing is crowding out US Treasury issuance, whether it is prompting investors to shift from corporate bonds to Treasuries, whether it is changing valuation comparisons across high-quality spread products, and whether it is crowding out debt issuance by other corporate sectors. The report concludes that current evidence is insufficient to support a clear crowding-out judgment, and that AI-related supply is more of a marginal pressure on long-end rate sentiment than a decisive driver.
Core views
The core views are: first, AI-related corporate bond supply has increased, but it differs from US Treasury issuance in term structure and market drivers, with the long end of Treasuries still mainly driven by fiscal, inflation, macro, and term premium factors. Second, the investor base and allocation constraints for corporate bonds and US Treasuries differ, so actual rotation is more likely to reflect changes in risk appetite than a pure supply shock. Third, IG corporate bonds compete more directly with high-quality spread products such as CLOs, RMBS, and MBS. Fourth, non-AI issuance has not been suppressed, and issuance by banks and other corporates remains active.
Analysis framework
The report analyzes supply-demand technicals, DV01 issuance size, term structure, intraday rate volatility, fund asset allocation, relative spread compensation, new issue concessions, secondary market spreads, and issuance forecasts, and combines these with views on relative value across USD/EUR IG, HY, CLOs, RMBS, leveraged loans, and macro rates.
Methodology notes
Assess whether supply is creating market pressure through issuance volume, investor demand, and new issue concessions.
The report distinguishes AI-related supply, non-AI supply, bank issuance, and overall IG/HY issuance, and observes whether concessions on non-AI new issues are widening to determine whether financing is being crowded out.
Use duration risk per basis point to measure the impact of different bond supply on long-end interest rate risk.
The report notes that the AI ecosystem has issued about $220 million/bp of DV01 this year, and compares IG issuance DV01 with US Treasury issuance DV01 to assess long-duration supply pressure.
Compare the yield compensation and technicals of IG corporate bonds with high-quality spread products such as CLOs, RMBS, and MBS.
The report argues that the key substitutes for IG corporate bonds are more likely to be high-quality structured products rather than US Treasuries; AAA CLOs and non-QM RMBS offer about 40bp of spread compensation relative to IG corporate bonds.
Assess the practical constraints on asset rotation through fund mandates, index composition, and return targets.
The report emphasizes that corporate bond-specific mandates and yield-oriented investors reduce the likelihood of large-scale rotation between corporate bonds and US Treasuries driven by supply shocks.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- USD IGOne of the core research subjects, most directly affected by AI-related long-duration issuance.
- Strengths
- Demand is supported by a high-yield environment, growth in corporate bond-specific mandates, and long-duration supply still holds value for liability-matching investors.
- Weaknesses
- Concentration of AI-related issuance has increased, heavy supply technicals may persist, and widening pressure in parts of TMT may spill over.
- Comparison
- Compared with US Treasuries, it is more driven by credit spreads and yield demand; compared with CLOs/RMBS, its spread compensation is less attractive.
- Risks
- Higher non-AI supply, AI-related headline risk, widening new issue concessions, and rising long-end rates.
- US TreasuriesUsed to assess whether AI corporate bonds are crowding out demand for or issuance of Treasuries.
- Strengths
- They remain the core asset for macro risk management and safe-haven allocation, with long-end yields mainly driven by fiscal and macro factors.
- Weaknesses
- The long end may face marginal sentiment pressure from AI-related corporate bond supply and hedging activity.
- Comparison
- The report argues that rotation from corporate bonds into US Treasuries is more likely to stem from declining risk appetite than from pure supply technicals.
- Risks
- Fiscal deterioration, inflation pressure, higher term premium, and long-end supply pressure.
- CLOsA high-quality spread substitute for IG corporate bonds.
- Strengths
- IG-rated CLOs offer higher spread compensation than IG corporate bonds, and EUR IG CLO tranches provide carry pickup relative to USD IG CLO tranches.
- Weaknesses
- The European macro and policy backdrop is relatively less favorable, and defaults in underlying assets may still create tail risk.
- Comparison
- AAA CLOs offer roughly a 40bp spread advantage relative to AA-rated IG corporate bonds.
- Risks
- Rising default rates, deteriorating liquidity, insufficient structural protection, and changes in refinancing conditions.
- RMBS / Agency MBSAlternative high-quality structured product allocations outside corporate bonds.
- Strengths
- Agency MBS offers hedging value, while non-QM RMBS provides about 40bp of spread compensation relative to IG corporate bonds.
- Weaknesses
- Non-QM RMBS carries additional credit risk, and its duration exposure does not fully match that of corporate bonds.
- Comparison
- Some structured products offer better compensation than IG corporate bonds without materially sacrificing credit quality.
- Risks
- Housing market volatility, prepayments, deteriorating credit performance, and interest rate volatility.
- EUR CreditA regional relative value comparison asset.
- Strengths
- Some EUR IG CLO tranches offer carry pickup relative to USD.
- Weaknesses
- Europe’s combination of growth, inflation, and monetary policy is relatively less favorable than that of the US, and the report expects EUR credit spreads to peak wider and recover more slowly.
- Comparison
- The report explicitly recommends Overweight USD vs. EUR.
- Risks
- European macro slowdown, policy uncertainty, and slower-than-expected recovery in credit spreads.
- USD HY / EUR HYA high-yield credit market comparison against the preference for IG.
- Strengths
- 2026YTD excess returns have been higher than IG.
- Weaknesses
- More sensitive to slower growth and default risk.
- Comparison
- The report prefers IG over HY, and within HY prefers BB, is Neutral on CCCs, and Underweight Bs.
- Risks
- Rising default rates, worsening rating migration, and persistently high financing costs.
Key data
- AI-related USD IG issuance volume$271 billionSince 2025, AI-related issuance excluding hyperscalers has entered the USD IG market.
- AI-related USD HY issuance volume$34 billionSince 2025, AI-related issuance has entered the USD HY market.
- Hyperscaler debt issuance$194 billionCovering IG markets including USD, EUR, GBP, CHF, CAD, and JPY.
- AI ecosystem DV01 issuance this year$220 million/bpThe report argues that due to its long-duration profile, DV01 better captures rate-risk supply than absolute issuance size.
- Leasing commitments of the five largest hyperscalersnearly $1.1 trillionBased on SEC filings, up $750 billion from early 2024, though this is a conservative upper bound for private bank financing of data centers.
- Annualized notional trading volume of OIS beyond 10 years$4.8 trillionUsed to compare the market’s capacity to absorb potential hedging demand.
- Non-AI issuance volume$1.29 trillionCombined IG and HY non-AI issuance this year, which the report says is the highest since 2020.
- USD IG spread forecast77bp current, 85bp 2026Q3, 82bp 2026Q4, 80bp 2027Q1, 78bp 2027Q2Updated as of 2026-07-15.
- USD HY spread forecast266bp current, 305bp 2026Q3, 295bp 2026Q4, 290bp 2027Q1, 280bp 2027Q2Updated as of 2026-07-15.
- EUR IG spread forecast77bp current, 91bp 2026Q3, 88bp 2026Q4, 86bp 2027Q1, 84bp 2027Q2Updated as of 2026-07-15.
- EUR HY spread forecast273bp current, 320bp 2026Q3, 315bp 2026Q4, 310bp 2027Q1, 305bp 2027Q2Updated as of 2026-07-15.
- Full-year USD IG gross issuance forecast$2.100 trillionFull-year 2026 forecast; 2026YTD is $1.411 trillion.
- Full-year USD HY gross issuance forecast$400 billionFull-year 2026 forecast; 2026YTD is $184 billion.
- Full-year EUR IG gross issuance forecast€850 billionFull-year 2026 forecast; 2026YTD is €527 billion.
- Full-year EUR HY gross issuance forecast€125 billionFull-year 2026 forecast; 2026YTD is €68 billion.
Impact & implications
For credit investors, the AI financing boom looks more like a marginal technical pressure on spreads and long-end rates than a systemic crowding-out of financing. The allocation implication is that investors should focus on comparing IG corporate bonds with high-quality structured products in terms of relative compensation, while also monitoring whether a rebound in non-AI supply triggers modest spread widening. Regionally, the report prefers USD over EUR; by rating, it prefers IG over HY, and seeks additional spread through BBB.
Risks
- AI-related issuance continues to expand and becomes concentrated among a small number of long-duration issuers, potentially worsening investor concentration and supply concerns.
- An increase in non-AI issuance could become the catalyst for the modest spread widening expected by the report.
- A deterioration in the US fiscal outlook, inflation pressure, or a rise in term premium could amplify long-end rate pressure.
- If macro downside risks increase, investors may shift from credit risk assets into US Treasuries.
- The relative value appeal of structured products may divert incremental capital away from IG corporate bonds.
- Recent widening in the TMT sector may transmit sentiment pressure to broader credit indices.
What to watch
- Whether the pace, term structure, and DV01 size of AI-related corporate bond issuance continue to rise.
- Whether new issue concessions for non-AI corporates and banks begin to widen meaningfully.
- Whether the ratio of USD IG issuance DV01 to US Treasury issuance DV01 rises further.
- Intraday moves in 30-year US Treasury yields and swap spreads on days of large AI or hyperscaler issuance.
- Fund flows into corporate bond-specific funds, ETFs, go-anywhere funds, and aggregate mandates.
- Changes in spread compensation of AAA CLOs, IG CLOs, Agency MBS, and non-QM RMBS relative to IG corporate bonds.
- The path of USD and EUR credit spreads, especially whether EUR spreads are wider and recover more slowly in H2 2026 as forecast.
- Whether non-AI issuance remains strong and whether M&A financing further increases supply.