25% of Europe's Corporate Debt Refis Within Two Years: The Corporate Layer, Not the Sovereign Curve, Is the Transmission Belt

The repricing is lopsided in a way the sovereign tape cannot show. Between 1 September and late October, European investment-grade yields rose about 30 basis points to 4.4%, while their spread over sovereigns widened only about 24 basis points to 117. High-yield borrowing costs rose roughly 90 basis points to 7%, with spreads widening 94 basis points to 354. The tail repriced three to four times faster than the index, because the tail is priced off the same sovereign curve plus a refinancing option that no longer exists.

On the index, the same episode looks mild. Amundi put euro investment-grade spreads near 93 basis points in mid-October after peaking at 100 on 30 September, and euro high-yield at roughly 330 against 260 at the end of August — a different benchmark from the bank's, but the same message: spreads stay orderly at the index level while the weak tail reprices violently, which is how the sovereign headlines and the corporate reality can both be true at once.

Below the index sits a set of borrowers for whom refinancing is no longer a question of price. Rating-agency work on Europe's ten most troubled borrowers identifies six with maturities landing in 2027 or 2028, and the weakest bonds are quoted at prices implying yields no lender would accept. The wall breaks in the high-yield tail rather than the investment-grade index.

Two features make this a wall rather than a list. First, concentration: European CLO portfolios held €5.3 billion of loans to CCC-rated companies maturing in 2028 at the end of June, up from €3.5 billion at the end of 2025 — a 51% increase in six months in exactly the bucket where recovery values are uncertain. Second, the pricing: Altice International's 3% January 2028 bond at roughly 56 cents implies a yield above 60%, and S&P Global Ratings estimates first-lien lenders recover only about 35% in default.

The evidence so far says the market has chosen extension over default, but not without a price. Arxada's English court approved a deal on 15 September pushing most of its debt about three years to 2031 and 2032; Colisée's April restructuring moved remaining debt to 2031 and was treated as a default before being upgraded to CCC in May; Merlin Entertainments secured new financing in early September covering its 2027 maturities, and its November 2027 bond improved to about 98 cents from roughly 95. Extension clears the maturity date and compounds the cost of the loan.

Rating agencies are treating the same pattern as a slow build rather than a cliff. Fitch Ratings raised its 2026 European leveraged loan default forecast to 4.0%–4.5% in July and kept the high-yield bond forecast at 3.75%–4.25%, while its October monitor recorded that European leveraged loan and high-yield default rates rose sharply in September as stressed Market Concern issuers transitioned to default. European default rates are turning up from a low base rather than from a crisis level.

Private credit is the buyer of last resort, and it is already charging for the job. Median first-lien loan gross spreads in Europe widened to about 632 basis points by June 2026 from 615 in March, against 554 basis points in the US, up from 525 in March. European first-lien loans priced 78 basis points wider than US first-lien loans in June, though part of that premium is compensation for leverage rather than geography: new European transactions closed at 5.12x total leverage against 4.62x in the US, with interest coverage near 2.5x on both.

The structure explains why the trade is not simply to buy the lenders. Apollo has published the thesis directly: non-bank lending is still only about 12% of European corporate lending against 75% in the US, and European direct-lending deals carry 25 to 50 basis points of extra spread over US equivalents. The five largest listed private-markets managers — Apollo, Ares, Blackstone, Carlyle and KKR — now manage a combined $1.5 trillion in perpetual capital. Europe remains the only large credit market where the private-capital build-out is still early enough to absorb a public-market failure.

This is where Europe's corporate problem reaches the United States, and the mechanism is not default correlation. It is the arithmetic of a shared investor base. Goldman Sachs' 23 October note projects hyperscaler bond issuance rising from $250 billion in 2026 to $420 billion in 2027, with global AI project finance potentially doubling to $300 billion. The European Central Bank has documented the other side of the same trade: the five largest US technology issuers — Amazon, Alphabet, Meta Platforms, Microsoft and Oracle — account for just under 10% of gross new euro-denominated non-financial corporate issuance, roughly €40 billion of hyperscaler bonds are outstanding, and hyperscalers drove 15% of the increase in euro-denominated corporate bond holdings in the year to March 2026. Every euro of AI paper competes with the euro a European corporate needs to refinance.

The crowding-out is visible in central bank market commentary. Demand for euro-denominated hyperscaler bonds stayed strong until mid-2026, then cover ratios began to edge down and credit spreads on hyperscaler paper started to widen while other issuers' spreads stayed flat — the first evidence that pressure is moving down the curve. US corporates have raised about €149 billion in European bonds in 2026, the second-highest annual total on record, and European high yield set a single-day issuance record on 16 September. Europe's corporate market has stopped being a European market and become a shared balance sheet.

The mechanism generalises, and it is measurable on the other side of the Atlantic. CoreWeave's five-year credit default swap spread blew past 1,100 basis points on 15 October, briefly above 1,140, and stood near 1,038 later in the session — 164% higher than a year earlier — while the yield to maturity on its 9.75% bond due 2031 reached 15.78%. Oracle's five-year credit default swap also hit record levels that week. A 1,000 basis point credit default swap quote marks the market's standard distress threshold, and CoreWeave crossed it while carrying more than $51 billion of debt and obligations.

Read the two stories together and the transmission becomes obvious. CoreWeave's debt rose $21.8 billion in six months while trailing revenue was $7.59 billion and trailing interest expense $1.87 billion, so roughly a quarter of every revenue dollar goes to interest — the same squeeze Goldman Sachs describes in Europe's weakest corporates, reproduced at the leveraged edge of the AI build-out. Europe's corporate layer and US investment-grade credit now resolve to one pricing problem across two currencies.

The first beneficiaries are the people paid to solve the problem, not the people who own it. Goldman Sachs reported third-quarter net revenues of $18.68 billion, with Global Banking and Markets at $13.46 billion (up 32% year on year), investment banking fees of $3.22 billion (up 21%), FICC at $3.47 billion, diluted earnings per share of $17.01 and a 12.4% standardized CET1 ratio on $101.4 billion of capital. JPMorgan Chase reported net income of $15.3 billion, earnings per share of $5.57, investment banking fees of $3.0 billion (up 14% year on year, down 8% sequentially), firmwide credit costs of $2.7 billion, net charge-offs of $2.2 billion and a 13.9% standardized CET1 ratio. Refinancing stress shows up as a fee event before it shows up as a loss event.

That ordering matters for how the story resolves. Banks are being paid handsomely to restructure, extend and re-paper Europe's weakest corporates — Goldman Sachs earns a spread for warehousing the paper that will not clear — while the loss is deferred to whoever holds the debt through the exchange, on bank and alternative-manager balance sheets. Banks earn fees to fix the wall in 2026 and hand the risk to whoever holds the exchanged paper in 2027.

The two horizons point in opposite directions, and only one of them is priced. Over the next two quarters the dominant effect is issuance and advisory volume plus the first maturity events of the wall. Over one to three years, the question is who absorbs the restructured paper. Fees arrive in 2026 and losses arrive in 2027.

The fact base is narrow and specific: European corporate borrowing costs have risen since September, the weakest borrowers are locked out of the public market, private credit has repriced the risk, and the same investor base is being asked to absorb a forecast $420 billion of AI paper in 2027. The inference is that Europe's stress now reaches US credit through capital scarcity and through restructured European paper sitting on bank and alternative-manager balance sheets, rather than through sovereign default correlation. The speculation is that this becomes systemic; Goldman Sachs does not expect that, and index-level spread evidence agrees so far. For investors, the trade that survives both outcomes is unglamorous: own the people paid to fix the wall before owning the people who own the damage, and treat 2027–2028 as the date to monitor rather than the date to front-run.