Thoughts on the Market

2026-07-17 · Hosted by Mike Wilson · Morgan Stanley

Executive Summary

Morgan Stanley's Lindsay Tyler (TMT Credit Research) and Anish Shah (Global Head of Debt Capital Markets) examined how credit markets are financing the AI infrastructure buildout, a scale Shah called "probably the most incredible time period" in his nearly twenty years at the firm. Equity research projects key hyperscalers could add over 30 gigawatts of data center capacity over two years, driving roughly $2 trillion in aggregate cash CapEx. Hyperscaler debt issuance has surged from under 1% to more than 10% of the investment-grade market, with Shah projecting total AI-related funding could reach 15% of all credit product issuance. The conversation covered innovation across investment-grade bonds, a new high-yield market for data center project finance (over $30 billion priced across 15 deals since fall 2025), and private credit/asset-backed structures, while flagging construction delays as the top risk to investor appetite.

Key Stories & Changes

1. AI Infrastructure Financing Reaching Unprecedented Scale

  • Morgan Stanley equity research projects key hyperscalers could add more than 30 gigawatts of data center capacity over a two-year timeframe, driving roughly $2 trillion in aggregate cash CapEx

  • A single gigawatt of data center capacity requires roughly $12 billion for the shell alone, and more than double that for chips and racks

  • Hyperscaler issuance has grown from less than 1% of the investment-grade bond market to more than 10% currently

  • Anish Shah projects AI-related funding (data centers and compute financing) could top 15% of total issuance across all credit products

  • Investment-grade bond volumes are likely to eclipse $2 trillion in issuance this year; $20 billion+ deals, once rare, now happen multiple times a quarter

2. New High-Yield Market for Data Center Project Finance

  • A first-of-its-kind high-yield bond structure has emerged to fund data center construction directly, bypassing traditional bank construction loans

  • More than $30 billion priced across 15 deals since fall 2025

  • Structure lets developers access fixed-rate, long-term capital while retaining the flexibility to call and refinance bonds at lower cost once construction is complete

  • Shah expects the innovation to expand further into other leveraged finance products given strong investor demand

3. Private Credit and Asset-Backed Financing Growth

  • Private credit has broadened from traditional leveraged buyout lending into financing large investment-grade-quality projects, including GPU/TPU-secured loans and data center development

  • Raises questions among investors about where risk ultimately resides across issuers, tenants, customers, and support providers — plus emerging concerns around circularity, vendor financing, and technology obsolescence risk

4. Credit Spreads Have Absorbed Supply Well

  • Despite the surge in issuance, investors have digested the supply "very well," with only modest spread widening for hyperscalers and tech issuers, described as "de minimis relative to their expected ROI"

1. AI Financing Is Reshaping Which Credit Markets Matter Most

The scale of AI-related capital needs is pulling financing activity across the full credit spectrum — investment-grade, high-yield, project finance, and private credit — in ways historically associated with capital-intensive utility, telecom, and energy buildouts, fundamentally changing how tech companies interact with debt markets after decades of minimal borrowing.

2. Innovation Concentrated in Sub-Investment-Grade Markets

While the bulk of issuance volume sits in investment-grade, Shah notes the bulk of structural *innovation* is happening in high-yield and private markets — new project-finance-style bonds designed specifically to meet AI infrastructure's unique construction-to-cash-flow timeline, a notable shift from how capital-intensive sectors were financed historically.

3. Return and Monetization Questions Remain Central to Credit Underwriting

Beyond simply funding capacity, investors are focused on how quickly new compute capacity translates into durable revenue, whether it supports proprietary products versus third-party leased compute, and how fungible that capacity is if demand patterns shift — credit work that is still evolving alongside the buildout itself. ---

Sentiment Analysis

Overall Market Sentiment: Constructive

Both speakers characterized credit market appetite for AI financing as strong and well-functioning, while flagging real, actively-monitored risks rather than dismissing them.

Risk Factors Highlighted

Construction delays: Named as the single biggest risk to investor appetite for AI-related debt; investors are doing heavy diligence and demanding structural protections.

Circularity and vendor financing risk: Private, asset-backed structures raise questions about interlinked financing arrangements among AI ecosystem participants.

Technology obsolescence risk: Rapid AI hardware evolution raises concerns about whether financed assets retain value over the life of the debt, even with amortizing structures.

Capacity fungibility risk: Uncertainty about whether data center/compute capacity can be redeployed across use cases (internal products vs. third-party leasing) if demand shifts.

Credit spread crowding: Heavy AI-related issuance could crowd out other areas of TMT credit issuance.

This episode was covered in today's [The Market Signal — 2026-07-17](https://marketsignal.beehiiv.com/p/the-market-signal-2026-07-17), a cross-source synthesis of multiple podcast reports.

Keep Reading