AI capex financing is shifting from pure debt to a mix of equity and convertibles

Overview

This report discusses why U.S. hyperscalers may increase equity and equity-linked capital in addition to investment-grade debt during the AI capex cycle. Barclays believes that training compute supply constraints, accelerating enterprise adoption, and improving model capabilities are jointly steepening the demand curve, pushing out the capex peak and making traditional capital intensity frameworks insufficient to fully explain current financing needs.

Core views

The core view is that the debt market is not closed, and hyperscalers do not currently face obvious near-term financing constraints; however, AI capex is characterized by long duration, scalable size, and still-uncertain ultimate monetization paths, so relying solely on fixed-maturity debt is not optimal. Equity and convertibles can diversify funding channels, preserve future debt capacity, strengthen credit metrics, and provide clearer downside buffers to credit investors. The report emphasizes that mandatory convertible preferred stock is a more precise tool for obtaining equity credit treatment and signaling balance sheet strength.

Analysis framework

Starting from capital structure and financing instrument selection, the report compares the roles of investment-grade debt, common equity, mandatory convertible preferred stock, zero-coupon convertibles, and other structured financing during the AI capex cycle. The focus is not on whether companies can raise financing, but on how they balance cost, dilution, credit metrics, issuance speed, and balance sheet flexibility across debt, equity, and convertible instruments.

Methodology notes

• Capital structure analysisAI capex financing mix framework

  Portfolio optimization across debt, equity, and equity-linked capital

  The report treats AI capex as a multi-year, long-duration investment cycle with uncertain return realization, so financing instrument selection needs to consider funding source diversification, future debt capacity, credit rating pressure, and common equity dilution simultaneously.

• Convertible strategyComparison of mandatory convertible preferred stock and zero-coupon convertibles

  Mandatory convertibles optimize equity credit treatment, while zero-coupon convertibles optimize coupon cost

  Mandatory convertible preferred stock typically carries higher equity credit, a clear three-year tenor, and a defined conversion path, making it better suited for signaling balance sheet strengthening; zero-coupon convertibles retain more debt-like characteristics, with their main advantages being lower cash coupon and higher conversion premium.

• Valuation and risk disclosureBarclays convertible bond valuation model

  Credit spreads and equity volatility are key variables in convertible bond valuation

  The disclosure section states that convertible bond valuation is based on Barclays' proprietary model, with key assumptions including credit spreads and equity volatility; any material change in either variable could significantly affect valuation.

Asset mapping & comparison

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

• GOOGL

  The core case in the report, using large-scale equity and mandatory convertible preferred stock to finance global compute infrastructure.

  Strengths

  Mandatory convertible preferred stock provides equity credit treatment, a defined conversion path, and a balance sheet strengthening signal, while preserving future debt capacity.

  Weaknesses

  Large-scale equity financing may trigger dilution debates and increase market focus on realizing returns from AI capex.

  Comparison

  Compared with zero-coupon convertibles, GOOGL's approach places greater emphasis on credit metrics and capital structure credibility rather than simply lowering coupon cost.

  Risks

  If AI demand, returns on compute investment, or subsequent financing needs fall short of expectations, market assessment of this financing path could be affected.

• ORCL

  A comparison case in the report, using mandatory preferred stock to support cloud and AI infrastructure financing plans.

  Strengths

  Equity instruments improve credit fundamentals, reduce ratings pressure, and provide clearer equity cushion to creditors.

  Weaknesses

  It still needs to prove that cloud and AI infrastructure investment can translate into sufficient cash flow.

  Comparison

  Similar to GOOGL, ORCL's mandatory instrument is viewed as proactive balance sheet strengthening rather than passive financing after debt markets have closed.

  Risks

  If capex rises faster than cash flow improves, credit metrics and market confidence could still come under pressure.

• META

  One of the potential follow-on issuers named in the report; the discussion is not whether it can raise financing, but how it balances debt, equity, and convertible instruments.

  Strengths

  It has the ability to finance AI capex, and if it adopts equity-linked issuance, that could reflect proactive balance sheet management.

  Weaknesses

  Additional equity-linked financing could bring dilution and investor communication pressure.

  Comparison

  The report groups META with MSFT and AMZN as hyperscalers that may follow the GOOGL/ORCL path.

  Risks

  If the market interprets financing as uncertainty over AI investment returns or funding pressure, valuation sentiment could be affected.

• MSFT

  A hyperscaler that could adopt a broader financing mix.

  Strengths

  It has ample funding channels and can choose among debt, equity, and convertible instruments to support long-term AI investment.

  Weaknesses

  Ongoing AI capex could increase pressure on capital allocation and cash flow.

  Comparison

  Like META and AMZN, the key issue is financing structure selection rather than financing availability.

  Risks

  The absorption capacity of investment-grade debt market supply, equity dilution concerns, and the pace of AI returns are key risks.

• AMZN

  A hyperscaler that could adopt a broader financing mix.

  Strengths

  It can support AI and cloud infrastructure buildout through diversified capital channels.

  Weaknesses

  Long-cycle capex could intensify the mismatch between free cash flow and investment needs.

  Comparison

  The report believes that, like META and MSFT, it has financing capability; the key is how it chooses capital structure instruments.

  Risks

  If the capex cycle continues to lengthen, debt capacity, credit investor confidence, and equity market acceptance will all need to be monitored.

• U.S. convertible bond market

  The primary market providing equity-linked capital for the AI value chain and large issuers.

  Strengths

  Fast issuance speed, with pricing possible in one day or overnight; flexible structures that can be paired with call spread and other dilution management tools.

  Weaknesses

  Valuation is sensitive to credit spreads, equity volatility, and the path of the underlying stock price.

  Comparison

  Compared with straight debt, convertible bonds provide equity upside; compared with common equity, they can defer or manage dilution.

  Risks

  If volatility, credit spreads, or equity market sentiment deteriorate, issuance windows and pricing attractiveness may decline.

• Neocloud and AI infrastructure issuers such as CRWV, NBIS, and IREN

  The report says these issuers have already shown broader adoption of convertible bond financing.

  Strengths

  Convertible bonds are well suited to innovative and growth industries raising capital during expansion.

  Weaknesses

  Their business models and cash flow maturity are usually weaker than those of hyperscalers.

  Comparison

  Compared with large companies such as GOOGL and ORCL, these issuers represent outer parts of the AI ecosystem that have relied on convertibles earlier.

  Risks

  If demand for AI infrastructure, financing conditions, or share price performance weakens, convertible bond refinancing and conversion prospects could come under pressure.

Key data

• Report date2026-06-09The cover date of the report is 9 June 2026.
• 2026 investment-grade bond issuance by hyperscalersMore than USD 200bn, with full-year expectations around USD 240bnThe report says this issuance volume already exceeds the annual issuance of the largest U.S. banks and represents an unprecedented level of supply for the non-financial sector.
• Total size of GOOGL equity financingUSD 84.75bnIncludes USD 16.75bn of three-year mandatory convertible preferred stock, USD 18bn of Class A and Class C shares, USD 40bn ATM, and a USD 10bn Berkshire Hathaway private placement.
• Size of GOOGL mandatory convertible preferred stockUSD 16.75bnThe report views this as a sign of expanding equity-linked capital within the AI capex financing toolkit.
• U.S. convertible bond issuance year to dateAbout USD 78bnThe report says this is the strongest start to a year since the index was created in 2003, above the same-period peaks in 2020 and 2021.
• Size of the U.S. convertible bond marketAbout USD 315.5bn increasing to about USD 471.5bnThe report says the current market value is at the all-time high for this asset class.
• Share of bond-like structures in the U.S. convertible bond marketAbout 85%The report notes that bond-style structures remain a common preference in the convertible market because they offer equity upside together with bond-like downside protection.
• Zero-coupon convertible bond examplesAMKR and ON issued 0% 2031 convertible notes in May 2026, with conversion premiums above 50%The report uses these cases to show that the market can absorb large zero-coupon convertibles, but their objective differs from that of mandatory convertible preferred stock.

Impact & implications

For investors, the AI capex financing narrative should not be viewed only through bond supply; attention should also be paid to how equity-linked capital is entering the capital structures of large technology companies. If more hyperscalers adopt mandatory convertibles or other equity-linked instruments, the market may interpret this as proactive maintenance of credit quality and investment capacity rather than simple funding pressure. But this also introduces new variables such as dilution, financing cadence, credit spreads, and valuation volatility.

Risks

• If enterprise AI adoption, model improvement, and the rollout of new use cases come in below expectations, returns on capex may fall short of financing assumptions.
• Training compute supply constraints and AI infrastructure buildout timelines may continue to delay the capex peak, widening cash flow gaps.
• The investment-grade debt market cannot expand indefinitely one-for-one with capex, so supply absorption capacity and spread changes need monitoring.
• Equity and mandatory convertible instruments may bring common equity dilution, investor acceptance risks, and market communication risks.
• Convertible bond valuation is highly sensitive to credit spreads, equity volatility, and the path of the underlying stock price.
• If financing is misread by the market as funding stress rather than proactive balance sheet management, valuation sentiment for related technology stocks could be hurt.
• The report discloses that Barclays and its affiliates may have business relationships with covered companies, and investors should use this report as only one factor in investment decisions.

What to watch

• Subsequent AI capex guidance from hyperscalers, the timing of the peak, and operating cash flow coverage capacity.
• Whether 2026 investment-grade bond issuance continues to approach or exceed the roughly USD 240bn full-year expectation.
• Credit metrics, ratings pressure, and market feedback following the GOOGL and ORCL mandatory convertible preferred stock transactions.
• Whether META, MSFT, and AMZN adopt equity-linked financing, and how the market interprets their motivations.
• Changes in issuance pace, market size, volatility, and credit spreads in the U.S. convertible bond market.
• The spread of convertible bond financing among neocloud and AI infrastructure companies such as CRWV, NBIS, and IREN.
• Shifts in issuance preference between zero-coupon convertibles and mandatory convertible preferred stock, especially the instrument choices of large technology issuers.