Global non-financial corporate debt crossed $100 trillion at the end of 2025 — the largest pool of private credit risk in human history. Corporate debt issuance hit a record $13.7 trillion in 2025, beating the 2021 pandemic-stimulus peak. Credit spreads sit near historic lows. And the rating agencies cannot agree on where defaults are heading: Fitch says falling toward 4.5%, Moody's says leveraged loan defaults could breach 7.9%. Underneath it all, nine AI hyperscalers need a projected $4.1 trillion in capital expenditure by 2030 — financed increasingly through the same bond markets this letter series has tracked since Letter 93. This is the most important credit story in the world that nobody is telling cleanly.
Not investment advice. Data sourced from OECD Global Debt Report 2026, IIF Global Debt Monitor February 2026, Fitch Ratings November 2025 Default Outlook, Moody's Leveraged Finance 2026 Outlook, S&P Global Ratings Default Transition and Recovery, BRG ThinkSet Maturity Wall Analysis. All figures current as of June 2026.
Global debt — government, corporate, and household combined — climbed to a record $348 trillion at the end of 2025, according to the Institute of International Finance's February 2026 Global Debt Monitor. Nearly $29 trillion was added in a single year, the fastest annual build-up since the pandemic surge of 2020. To appreciate the scale: that one-year increase alone is larger than the entire annual GDP of the United States. Within that total, non-financial corporate debt — money owed by operating businesses, not banks, not governments, not households — reached approximately $100.6 trillion, the largest single component of the entire $348 trillion stockpile, ahead of government debt at $106.7 trillion and well ahead of household liabilities at $64.6 trillion.
This letter series has spent significant ground covering sovereign debt (Letter 99), bond markets (Letter 98), and private credit (Letter 93). Corporate debt — the conventional, public bond and syndicated loan market that dwarfs all of them combined in absolute terms — has not yet received its own dedicated treatment. That gap closes here. The OECD's Global Debt Report 2026 puts outstanding global corporate debt securities and syndicated loans at $59.5 trillion as of year-end 2025 — $36.4 trillion in corporate bonds, $23.1 trillion in syndicated loans — with the IIF's broader $100.6 trillion figure capturing additional bank lending and other forms of corporate borrowing beyond the OECD's bond-and-loan-market scope. Either measure describes the same underlying reality: corporate debt is now the largest, fastest-growing, and least-discussed pool of credit risk in the global financial system.
2025 was, by the OECD's own characterisation, the strongest year for corporate debt markets on record. Companies globally borrowed $13.7 trillion across corporate bond and syndicated loan markets — $6.8 trillion in bonds, $7 trillion in syndicated loans — surpassing the previous 2021 peak, which had itself been driven by emergency pandemic-era financing needs and extraordinarily loose monetary policy. This time the loose-money explanation does not fully apply: rate cuts across the US, UK, and eurozone through 2025 helped, but the scale of the increase reflects something structurally different — genuine, broad-based corporate demand for capital, not crisis-driven emergency borrowing. LSEG separately recorded global debt issuances (including financial-sector and sovereign issuance) of $12.1 trillion for 2025, up 13% year-on-year, with 36,000 individual offerings brought to market — also an all-time record.
The pricing tells an equally striking story. Corporate credit spreads sit near historical lows for both investment-grade and non-investment-grade issuers, despite the record borrowing volumes, despite geopolitical tensions running through nearly every letter in this series, and despite tariff-driven trade uncertainty. The OECD attributes this partly to genuinely strong corporate fundamentals — high cash levels, solid aggregate credit quality, default rates below historical averages — and partly to factors unrelated to corporate credit quality at all, a deliberately understated way of flagging that technical market dynamics (yield-starved investors chasing any spread above government bonds, the same liquidity wave examined in Letter 83) are also compressing spreads independent of whether the underlying credit risk has genuinely improved.
The single most important analytical fact in global corporate credit right now is one that gets remarkably little attention: the major rating agencies do not agree on whether defaults are rising or falling, and the gap between their forecasts is not small. This is not a rounding-error disagreement about methodology. It reflects genuinely different underlying assumptions about how the leveraged loan market — the riskiest, most private-credit-adjacent segment of corporate debt — is actually behaving.
S&P's October 2025 data sits closer to Fitch — a 12-month trailing speculative-grade default rate of 4.4% in the US and 3.50% in Europe, "well below 2024 levels" — but flags that distressed exchanges (companies restructuring debt outside formal bankruptcy, swapping bondholders into worse terms to avoid a formal default classification) continue to account for a large share of year-to-date defaults. This matters enormously: Moody's own research shows the high-yield default rate is roughly one-third lower when distressed exchanges are excluded from the calculation — meaning a meaningful share of what counts as "low defaults" in some methodologies is actually companies restructuring under duress in a form that doesn't trigger the official default label. Whether you count distressed exchanges as defaults is not a technicality. It is the difference between Fitch's optimistic 2.8% and a materially higher real number.
The leveraged loan market — debt extended to companies that already carry significant existing debt or weak credit profiles, predominantly used to finance private equity buyouts — is where the genuine stress concentrates, and it is also the segment most directly connected to the private credit thesis this letter series examined in Letter 93. Moody's flags explicitly that the most distress is concentrated among "borrowers with loan-only capital structures backed by private equity firms" — precisely the companies whose debt was extended by, or now competes directly with, the private credit funds (Blackstone, Apollo, Ares, Blue Owl) covered in that earlier letter.
The structural decay from 2008 was never fixed. Covenant-lite loans — lacking the maintenance tests that traditionally let lenders intervene early when a borrower's finances deteriorate — reached 77.4% of US leveraged loans by mid-2018 and have stayed structurally elevated since. The Bloomberg assessment from 2020 still holds: "the aggressive push toward weakening protections virtually ensures that recovery rates will be worse" whenever the credit cycle genuinely turns, regardless of which agency's default forecast proves correct.
Moody's explicitly flags that competition between syndicated leveraged loans and private credit (Letter 93) is weakening covenants and adding leverage risk industry-wide, as both channels compete for the same borrowers by offering progressively looser terms. CLO structures will help absorb defaults, but tight loan spreads limit managers' ability to actively optimise portfolios — meaning the buffer exists but is thinner than headline CLO resilience suggests.
Chinese corporate debt issuance has driven a disproportionate share of global emerging-market corporate borrowing, alongside Brazil and Korea. China became one of the largest corporate bond markets in the world in under two decades — Chinese corporate bonds grew from $69 billion in 2007 to $2 trillion by 2017, and have continued expanding since, a trajectory this letter series first flagged in the reindustrialisation analysis (Letter 97).
Despite the European rearmament wave and NATO's 2025 commitment to push defence spending toward 5% of GDP by 2035, the IIF notes that defence-sector firms — particularly SMEs in the supply chain — have had limited ability to tap corporate debt markets, even with strong investor demand. This is a genuine financing gap connecting directly to the reindustrialisation thesis (Letter 97) that has not been resolved by the broader 2025 issuance boom.
The single most consequential development reshaping global corporate debt markets right now is not a default cycle or a maturity wall in the conventional sense — it is the financing requirement of the AI buildout. The OECD's Global Debt Report 2026 puts the projected capital expenditure of nine major AI hyperscalers — primarily US technology firms — at $4.1 trillion between 2026 and 2030. For comparison: total capital expenditure by all non-financial companies in the United States in 2025 was just above $3 trillion. Nine companies' five-year AI infrastructure spending plan exceeds an entire year of capital investment by every other US company combined.
OECD Global Debt Report 2026 · What happens if hyperscalers fund this through bond markets
The cumulative hyperscaler bond issuance of $122 billion in 2025 represented no more than 15% of total investment-grade issuance by non-financial US firms — absorbed comfortably, "without market-wide friction," in the OECD's words, though credit risk metrics of individual issuers have sometimes responded sharply to large single deals. The genuine question is what happens as the $122 billion figure scales toward a meaningful fraction of $4.1 trillion over the next five years. If even half of future hyperscaler capex needs are financed through bond markets, the OECD calculates that borrowing by these nine issuers alone would amount to an average of 15% of historical global gross corporate issuance, every year, for half a decade — a level of single-sector concentration in global credit markets that has no recent precedent outside of sovereign or banking-sector borrowing.
Collateralised loan obligations — securities backed by pools of leveraged loans, sliced into tranches of varying risk — remain the primary structural mechanism absorbing leveraged loan default risk. Moody's 2026 outlook is explicit: CLO structures will help absorb defaults even as covenant-lite issuance keeps rising, but tight loan spreads will limit managers' ability to actively optimise portfolios — meaning the structural buffer exists, but the managers running it have less room to manoeuvre defensively than in prior cycles, making credit monitoring "essential" rather than routine.
The same firms covered in Letter 93's private credit analysis are now a direct structural force shaping syndicated corporate debt market terms, not merely a parallel, separate channel. Moody's flags the competitive dynamic explicitly: private credit's growth is weakening covenants in the syndicated market as both channels race to offer borrowers looser terms. The Q1 2026 private credit stress examined in Letter 93 and the corporate debt dynamics in this letter are not two separate stories — they are increasingly one story, viewed from two different entry points.
Microsoft, Amazon, Google, Meta, and their peers were not, until very recently, significant corporate bond issuers relative to their balance sheet strength — most funded growth from operating cash flow. The scale of AI infrastructure investment is changing this fundamentally, and these issuers carry investment-grade credit profiles that make their debt highly attractive to yield-starved fixed income investors — the same liquidity dynamic from Letter 83 — even as the underlying capex commitment scales toward a size with no recent precedent for a single sector's claim on global bond market capacity.
Mutual funds, including the high-yield specialist funds whose assets under management roughly doubled in the decade before 2019, remain a significant structural holder of corporate credit risk — including leveraged loans, an asset class with daily redemption-eligible fund structures wrapped around fundamentally illiquid underlying loans. This liquidity mismatch — open-ended funds holding illiquid leveraged loans — is the retail-market equivalent of the private credit BDC liquidity mismatch examined in Letter 93's analysis of the Q1 2026 gating crisis.
"Record issuance at near-historic-low spreads" is not unambiguously good news, and the OECD's own framing — corporate debt markets "becoming more like equity markets" — is a genuine warning, not a neutral observation. When debt markets price risk the way equity markets do (chasing momentum and narrative rather than discounting fundamental credit quality with discipline), the entire purpose of credit markets as a risk-pricing mechanism degrades. Spreads near historic lows, despite $13.7 trillion in record issuance and despite geopolitical tensions running through nearly every other letter in this series, is the kind of disconnect between price and risk that historically resolves through a sudden, rather than gradual, repricing.
The rating agency disagreement on default trajectory is not a minor technical dispute — it reflects a genuine, unresolved disagreement about whether distressed exchanges should count as defaults, and that classification choice changes the headline number by roughly a third. Investors relying on a single agency's "defaults are falling" narrative without understanding the distressed-exchange methodology question are working from an incomplete picture. The honest state of the world, as of June 2026, is that nobody — including the rating agencies whose job is precisely this assessment — has full confidence in where corporate default rates are actually heading over the next twelve months.
The AI capex financing wall is the single largest unpriced structural risk in global corporate credit markets, and almost no commentary outside specialist OECD and BIS publications is treating it with the weight it deserves. $4.1 trillion in five-year capital expenditure commitments from nine companies, an average 15% annual claim on historical global gross issuance if half-debt-financed, and credit markets that have so far absorbed early hyperscaler issuance "without market-wide friction" — this is either evidence that global bond markets have ample capacity to absorb genuinely productive capital investment, or it is the early stage of a single-sector concentration risk that has no recent precedent and that current spread pricing is not yet reflecting. This letter does not resolve that question. It is the question.
Long-horizon thinking on capital, technology, and the forces shaping the next decade of wealth creation. Written from first principles. Not consensus. Not noise.