The AI financing cycle is reshaping the USD IG credit curve

Overview

This report focuses on spread curves, regional relative value, and primary issuance structure in global credit markets. The core view is that the AI-related capex and data-center financing cycle has had a visible impact on the USD IG market: heavier long-dated bond supply from Technology/TMT has made the TMT 10s30s curve materially steeper relative to non-TMT, while long-end supply remains limited outside Technology and Utilities, so this curve divergence may persist. Regionally, the report has a modest preference for USD credit over EUR credit, but also points out that rising cross-border issuance is blurring traditional regional risk classifications.

Core views

First, the USD IG 10s30s curve appears to be near the flatter range on the surface, but internal structure is clearly diverging, with TMT steepening on AI-related long-term financing while other non-financial sectors are flattening. Second, changes in the front-end 2s10s and 5s10s curves are more influenced by the U.S. Treasury yield curve and the absolute yield environment; if the Treasury curve continues to steepen within the year, front-end spread curves in non-financial IG could move back toward the flatter levels seen at the start of the year. Third, the EUR IG 5s10s curve is relatively flat and shaped by technical factors, limiting the relative value of extending duration, while the 5-7 year tenor is more attractive. Fourth, the rising issuance and index share of North American companies in the EUR IG and EUR HY markets make the delineation of regional credit exposure more complex.

Analysis framework

The report mainly analyzes USD and EUR credit markets through issuance maturity structure, issuer-matched credit spread curves, index spread levels, yield-curve shape, primary market supply, cross-border issuance share, and index market-value composition. The analysis also incorporates Goldman Sachs' forecast panels for spreads, returns, issuance, defaults, and rating migration to validate its relative-value views.

Methodology notes

• Curve AnalysisIssuer-Matched Spread Curves

  issuer-matched spread curves

  The report pairs outstanding bonds from the same issuer, with the same seniority but different maturities, to reduce distortions in unmatched curves caused by composition differences.

• Relative ValueConditional Comparison of Curve Slope and Index Spread

  curve slope versus index spread level

  The report դիտches slopes such as 2s10s together with current index spread levels, and argues that in a tight-spread environment with IG OAS around 75bp, extremely flat curves are not unusual relative to history.

• Supply-Demand TechnicalsPrimary Issuance Maturity Structure Analysis

  gross issuance by tenor and sector

  The report uses issuance shares across maturities, sectors, and parent-company domicile to assess whether long-end supply is sufficient to alter curve shape.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• USD IG

  One of the report's core analytical focuses, with emphasis on the 10s30s, 2s10s, and 5s10s spread curves.

  Strengths

  It has modestly outperformed EUR IG year to date, and yield-based demand still supports certain maturities.

  Weaknesses

  Overall spreads are already tight, and the front-end curve could flatten again if the U.S. Treasury curve steepens.

  Comparison

  The report has a modest preference for USD credit over EUR credit.

  Risks

  If long-end supply broadly rebounds due to M&A financing, curves and long-dated bond performance could come under pressure.

• TMT / Technology USD IG long-end bonds

  The sector directly affected by the AI-related financing cycle.

  Strengths

  Demand for long-dated bonds is still supported by yield-based buyers.

  Weaknesses

  AI-related issuance has longer maturities, and concentrated supply is pushing up the steepness of the TMT 10s30s curve.

  Comparison

  Relative to non-TMT non-financial sectors, the TMT curve has steepened markedly since mid-2025.

  Risks

  Continued expansion of financing for data centers and AI projects could further increase long-end supply.

• EUR IG

  A regional credit market emphasized in the report's comparisons.

  Strengths

  The 5-7 year tenor segment is considered relatively more attractive.

  Weaknesses

  The curve is relatively flat, and compensation for extending duration is insufficient.

  Comparison

  It is slightly less attractive than USD credit, though tenor selection within the market remains important.

  Risks

  European macro divergence, commodity-flow disruptions, and technical supply changes could affect the curve.

• EUR HY

  Used to observe cross-border issuance by North American issuers and regional exposure in the high-yield market.

  Strengths

  Issuance by North American companies is active, and market depth and issuer composition are becoming more diversified.

  Weaknesses

  Regional labels do not fully align with the true sources of issuer risk.

  Comparison

  North American issuers account for a high share of issuance in EUR HY, showing that the European high-yield market is increasingly influenced by U.S. companies.

  Risks

  Rising cross-border issuance could make regional risk, FX, and funding-cost exposures harder to manage.

Key data

• Weighted average maturity of AI-related IG issuance14.8 yearsHigher than the roughly 12.8-year weighted average maturity of non-Technology non-financial IG supply.
• Technology share of USD IG long-end supply beyond 25 years31%Technology is the largest source of long-end supply this year.
• Announced but uncompleted global strategic M&A backlog$1.1 trillionThe report believes this could bring some debt financing over the medium term, but it has not yet translated into broad 30-year supply.
• USD IG OASabout 75bpThe report uses this level to explain the reasonableness of the currently flat 2s10s curve.
• Historical percentile of the USD IG non-financial 2s10s curve11th percentileThe absolute level is extremely flat, but it is closer to fair value in a tight-spread environment.
• Historical percentile of the EUR IG 5s10s curve33rd percentileRelative to the historical range since 2015, it has returned to a more normal range.
• Share of EUR IG 10-year supply23%Below 29% in 2024, one of the technical factors behind the flattening of 5s10s.
• Share of EUR IG 2-4 year supply20%Above 16% in 2024.
• North American issuers' share of EUR IG index market value18.4%The highest since 2020.
• North American issuers' share of EUR HY index market value8.2%The highest since 2022.
• Year-to-date share of total EUR IG issuance contributed by North American parents19.4%Close to or above 2025 levels, representing a relatively high contribution since 2016.
• Year-to-date share of total EUR HY issuance contributed by North American parents28.8%Close to the high end of the historical range.
• Year-to-date total EUR IG issuance by North American companies€98 billionAs of 2026-07-01, Financials and Technology together contributed 70%.
• EUR IG Technology issuance by North American companies€33 billionAlready above most full-year levels since the global financial crisis.
• Total EUR HY issuance by North American companies€18 billion2026 still has the potential to be one of the more active years for this type of supply.

Impact & implications

For portfolios, the report suggests investors should not look only at overall USD IG or EUR IG curve levels, but should break them down by sector, tenor, and issuance source. AI financing is pressuring the long end of TMT and creating curve divergence, while insufficient non-TMT long-end supply supports relative performance at that end. In regional allocation, USD credit has a slight edge over EUR credit, but as North American issuers increasingly enter the EUR market, investors need to reassess the true regional exposure behind index labels.

Risks

• If M&A financing translates into broader 30-year bond supply over the medium term, it could weaken the report's view that non-TMT long-end supply is insufficient.
• If the U.S. Treasury yield curve evolves differently from Goldman Sachs' rates team's forecast, front-end credit spread curves could deviate from expectations.
• If AI-related capex and data-center project financing continue to accelerate, supply pressure at the long end of Technology/TMT could rise further.
• The EUR credit market is heavily influenced by European macro divergence, central-bank reaction functions, commodity-flow disruptions, and the growth-inflation mix.
• Rising cross-border issuance means regional index labels do not fully match the true sources of issuer risk, potentially leading to portfolio exposure misjudgment.

What to watch

• The scale, tenor, and sector concentration of new issuance related to AI and data-center project financing.
• Whether the Technology and non-TMT non-financial USD IG 10s30s curves continue to diverge.
• Whether the U.S. 2s10s and 5s10s Treasury yield curves continue to steepen within the year.
• Changes in the shares of EUR IG 10-year and 2-4 year supply, and whether the 5s10s curve continues to be flattened by technical factors.
• The index market-value share and primary issuance share of North American companies in EUR IG and EUR HY.
• Whether announced M&A transactions convert into debt financing and long-dated bond issuance.
• Changes in spreads, excess returns, default rates, and rating migration in Goldman Sachs' forecast panels.