AI financing demand is large, and a single public bond market cannot support it alone
AI summary card
AI financing demand is large, and a single public bond market cannot support it alone
Goldman Sachs expects the five largest hyperscalers to spend a total of $5.8 trillion in AI capital expenditure between FY2025 and FY2030, and believes financing needs will ultimately require IG, HY, loans, private infrastructure, project financing, bank financing, and multi-currency markets to work together.
- USD IG issuance year-to-date in 2026 has reached $1.37 trillion, well above the pace of recent years, increasing the upside risk to the full-year $2.1 trillion forecast.
- The five hyperscalers have issued approximately $194 billion across currencies so far this year, accounting for about 9% of total IG supply; 32% of their combined IG issuance is priced outside the USD market.
- Raising the five hyperscalers’ index-eligible debt to the level of large U.S. banks implies about $510 billion of incremental debt capacity, but may create index-weight and concentration constraints.
- In the private market, the infrastructure bucket has financed over $140 billion of data center deals since early 2025, and Goldman expects the infrastructure asset class to exceed $3.0 trillion by 2030.
- The bank sector is seen as a core holding across scenarios; Goldman has broadened its constructive stance on USD and EUR bank credit and is more willing to step down into Tier 2 within EUR banks.
Report interpretation
Overview
This report discusses the impact of AI-related financing demand on global credit markets and extends to evaluate bank sector performance and positioning value before and after 2Q2026 earnings. The core conclusion is that the AI capital expenditure cycle is enormous, and although the USD IG corporate bond market is deep, if it were to absorb the majority of financing, it would push hyperscalers’ index weights higher and create concentration constraints. In coming years, financing needs will require public credit markets, private infrastructure, project-finance-style joint structures, bank financing, and multi-currency issuance to participate together.
Core views
First, AI-related debt supply remains elevated, with Technology and adjacent sectors driving USD IG issuance year-to-date in 2026 to $1.37 trillion, and if the summer slowdown does not materialize, upside risk to the full-year issuance forecast increases. Second, low leverage at hyperscalers leaves room for incremental debt, but investor limits on issuer concentration, broad technology exposure, and duration structure may form a cap. Third, data center and AI infrastructure financing should continue to spill over into private infrastructure, loans, HY, project financing, and bank financing. Fourth, bank credit fundamentals remain supported, and Goldman views the bank sector as a core holding, expanding its constructive view on USD and EUR bank debt.
Analysis framework
The report uses primary market issuance data, cross-currency IG supply, hyperscaler debt-capacity estimation, large U.S. banks as a concentration benchmark, private data center financing scale, and USD/EUR bank credit performance as its main analysis framework, while incorporating fundamentals, technicals, valuation, capital structure, and regional fiscal-risk premium in relative-value assessment.
Methodology notes
Evaluates AI financing pressure by using year-to-date totals, sector share of issuance, monthly issuance pace, and cross-currency hyperscaler issuance.
The report compares 2026 USD IG issuance, June and early July issuance pace, Technology sector share, and the share of hyperscalers in total IG supply to gauge absorption capacity in the public debt market.
Uses debt size and index weight of large U.S. banks in the Bloomberg USD IG Corporate Index as a reference ceiling for potential hyperscaler issuance.
Raising index-eligible debt of the five hyperscalers to the level of large U.S. banks results in about $510 billion of incremental capacity, but the report notes this comparison may be optimistic.
Compares performance and spreads between USD banks’ Money Centers, Regionals, and Yankees, and among EUR banks’ senior, Tier 2, AT1, and EUR HY.
Combining fundamentals, supply, duration, fiscal risk premium, and recent performance, the report broadens the constructive bank view and favors Tier 2 over senior within EUR banks.
Asset mapping & comparison
Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).
- USD IG corporate bondsOne of the main public market channels absorbing AI financing
- Strengths
- Large market size, strong liquidity, and ability to support large investment-grade issuance.
- Weaknesses
- If this market is relied upon excessively, hyperscaler weights in indices may rise significantly above current market norms.
- Comparison
- Compared with private and bank financing, USD IG is more transparent and larger in scale, but compared with a multi-currency, multi-tool mix, concentration risk is higher.
- Risks
- Issuer concentration, market saturation, long-duration supply pressure, and limits on investor appetite for broad technology exposure.
- EUR IG corporate bondsAI-related cross-currency financing and bank credit positioning market
- Strengths
- Diversifies pressure away from the USD market and offers relative-value opportunities in bank credit.
- Weaknesses
- More visibly affected by regional macro factors and fiscal-risk premia.
- Comparison
- EUR banks’ performance is closer to the broader market than USD banks, with stronger relative year-to-date excess return.
- Risks
- France and UK banks may continue to carry spread premia versus peers due to sovereign fiscal concerns.
- Private infrastructure creditAn important incremental channel for data center and digital infrastructure financing
- Strengths
- Well suited for long-dated, project-style, asset-backed data center financing.
- Weaknesses
- Usually lower transparency and lower liquidity than public bond markets.
- Comparison
- Compared with public IG bonds, private infrastructure is more suitable for customized and project-based financing.
- Risks
- Project execution, data center demand, financing costs, and asset valuation risk.
- High yield bonds and leveraged loansSupplementary financing channels for the AI ecosystem and related investment chains
- Strengths
- Can serve non-investment-grade or higher-leverage issuers, expanding financing capacity.
- Weaknesses
- Higher funding costs and greater sensitivity to risk appetite and default cycles.
- Comparison
- Compared with IG bonds, HY and loans cover a wider credit-quality spectrum but have higher volatility and credit risk.
- Risks
- Widening spreads, rising defaults, declining liquidity, and reversals in investor risk appetite.
- Banks sector creditCore credit allocation segment in the report
- Strengths
- Fundamental support, relatively manageable supply versus market size, and shorter duration than non-bank and technology issuance.
- Weaknesses
- Sub-sector and capital-structure performance is divergent, and issuance pace can still affect relative outcomes.
- Comparison
- USD banks had previously lagged the broader USD IG market slightly; EUR banks performed closer to or somewhat better than broader EUR IG.
- Risks
- Unexpected post-earnings supply in banks, recurring macro and geopolitical risks, and sovereign fiscal-risk premia in France and the UK.
Key data
- USD IG total issuance year-to-date in 2026$1.37 trillionHigh issuance pace from Technology and adjacent categories has pushed supply well above recent years.
- 2026 USD IG full-year issuance forecast$2.1 trillionThe report believes the risk is clearly tilted higher, especially if the summer slow-down does not occur.
- 2026 June USD IG issuance$240 billionWell above the 2020-2025 June average of $129 billion.
- July 2026 USD IG issuance as of July 8$51.5 billionA faster start versus the 2020-2025 July average of $104 billion.
- Technology share of USD IG issuance year-to-date in 202620%A new high in Dealogic data, and potentially understates true momentum due to AI ecosystem classification differences.
- Five hyperscalers' expected AI capital expenditureTotal FY2025-FY2030 $5.8 trillionInvolves Microsoft, Amazon, Meta, Alphabet, and Oracle.
- Five hyperscalers’ cross-currency debt issuance this yearabout $194 billionAbout 9% of total IG supply, with 32% of the combined IG issuance priced outside the USD market.
- USD IG corporate bond index size$7.9 trillionMarket depth is substantial, but issuer concentration and index-weight effects may still be constraining.
- Estimated incremental debt capacity for hyperscalersabout $510 billionBased on a simple estimate of lifting their index-eligible debt to the level of large U.S. banks.
- Private data center financingOver $140 billion since early 2025Includes only data center transactions, excluding other uses such as chips.
- Potential infrastructure asset-class sizeOver $3.0 trillion by 2030Includes the infrastructure asset class, including digital infrastructure.
- USD IG banks year-to-date excess return47bpBelow the broader USD IG market's 68bp.
- EUR IG banks year-to-date excess return81bpAbove the broader EUR IG index's 73bp.
- USD IG bank issuance$519 billionUp 25% year-over-year, but around 40% of overall USD IG supply and at the lower end in recent years.
- EUR IG bank issuance€225 billionBroadly in line with last year, around 44% of overall EUR IG supply.
Impact & implications
AI financing demand may continue to reshape credit-market supply structure and relative-value dynamics. For investors, the key is not simply whether hyperscalers can lever up further, but whether the public debt market can actually absorb that scale under issuer concentration, duration, sector exposure, and investor risk-budget constraints. Multi-channel financing will create incremental opportunities for private infrastructure, data center financing, project financing, and bank financing, while also requiring public credit investors to price AI-related issuance more precisely. For bank exposure, despite heavy supply, conditions remain relatively manageable and fundamentals are still supported, so the bank sector can remain a core credit holding; however, EUR banks still require separation of sovereign fiscal-risk premium by country.
Risks
- AI-related issuance continues above expectations, increasing supply pressure and issuer concentration in the USD IG market.
- If the summer issuance slowdown does not materialize, 2026 full-year USD IG issuance could finish materially above forecast.
- Increased long-duration debt supply from hyperscalers weakens investor absorption capacity.
- Investors already carry large Technology exposure, which may limit willingness to add hyperscaler debt allocation.
- Private infrastructure and project financing may fail to absorb data center financing demand as expected.
- Unexpected post-earnings bank supply in 2Q2026 drags relative performance.
- EUR financial institutions face persistent spread differentiation due to French and UK fiscal-risk premia.
What to watch
- Whether USD IG issuance in summer 2026 actually slows.
- Changes in the share of Technology and AI-related issuance in total USD IG supply.
- Cross-currency issuance scale and maturity structure of hyperscalers such as Microsoft, Amazon, Meta, Alphabet, and Oracle.
- Changes in hyperscalers’ weights in the Bloomberg USD IG Corporate Index.
- Incremental scale of data center private financing, project-finance-style joint structures, and bank financing.
- USD and EUR bank issuance pace after 2Q2026 earnings.
- Spread divergence among USD bank Money Centers, Regionals, and Yankees.
- Relative performance between EUR bank senior, Tier 2, AT1, and EUR HY.
- Fiscal-risk premium of French and UK banks versus European peers.