Riskiest US Firms Squeezed as Treasury Yields Near 4.8%

What happened

A sharp rise in U.S. Treasury yields is rippling through the corporate bond market, and the pain is landing hardest on the weakest borrowers. The 10-year Treasury yield climbed to roughly 4.82% this week, its highest level since January 2025, while the 30-year yield pushed above 5.3%, territory last seen around 2007 11017. At the same time, spreads on junk-rated corporate debt — the extra yield investors demand over Treasuries to hold risky bonds — have widened to their highest levels since the market turmoil that followed last year's tariff-driven 'Liberation Day' selloff, according to the Financial Times 1119.

That combination matters because corporate bonds are priced as a Treasury yield plus a credit spread. When both move higher at once, the effect compounds: companies refinancing maturing debt or borrowing fresh capital face a double hit, paying more simply because the government benchmark rose and paying even more because investors are pricier about risk 41213.

The move is not confined to the U.S. Yields in Japan, Germany, France and the U.K. have all pushed to multi-decade highs, with Japan's 10-year bond touching 3% for the first time since 1996 and U.K. 30-year gilts at their highest since the late 1990s 161837. Analysts describe this as fundamentally a global story, driven by heavy government borrowing, resilient growth, inflation risk tied in part to oil prices, and doubts about whether central banks will cut rates as quickly as once expected 2518.

Why it's happening

Reuters' explainer lays out the mechanics clearly: investors are demanding more compensation to hold government debt because of a flood of new issuance, sticky inflation, geopolitical energy risk, and uncertainty about the Federal Reserve's next moves, compounded by questions over whether foreign buyers still want as much U.S. debt as before 512. U.S. government debt has surpassed $40 trillion, and rising yields raise the cost of servicing that load, creating a feedback loop in which fiscal worry pushes yields higher, which then worsens the fiscal picture 168.

A second, more novel driver has emerged this year: artificial intelligence. The five major hyperscalers — Alphabet, Amazon, Meta, Microsoft and Oracle — have issued roughly $200–220 billion in debt so far in 2026 to fund data centers and AI infrastructure, more than double last year's pace and a sharp break from an era when these companies mostly self-funded from cash flow 171612. The New York Times traces the escalation precisely: hyperscaler debt issuance averaged under $30 billion a year from 2020 through 2024, jumped past $100 billion in 2025, and has already topped $200 billion this year 17. Oracle, the lowest-rated of the five, has seen its borrowing spread over Treasuries widen from 1.05 percentage points to 1.45 points in a matter of months — a sign that even blue-chip tech names are paying more, though not because anyone doubts their solvency 17.

That AI borrowing wave is squeezing weaker companies indirectly, by competing for the same pool of investor capital and adding to the sheer volume of debt markets must absorb 1217. Analysts are split on how much AI debt is pushing up Treasury yields specifically versus corporate spreads more narrowly — a distinction addressed below.

Who is actually at risk

The research is consistent that this is not a uniform corporate crisis but a fragmenting one. PIMCO's credit analysis finds that median interest-coverage ratios remain solid — around 6 times for investment-grade issuers and 3 times for high-yield issuers — suggesting most companies can absorb higher refinancing costs 41315. The exception is CCC-rated borrowers, the bottom rung of junk credit, where PIMCO estimates coupons on bonds maturing in 2027 and 2028 could roughly double if refinanced at today's yields 41315.

CNBC's reporting broadens the risk list to small-cap companies carrying floating-rate debt, along with commercial real estate, private-equity-backed firms, direct-lending portfolios and weaker software businesses — all sectors financed on the assumption that cheap capital would persist indefinitely 18. Teneo's consulting analysis adds a scale dimension: more than $1.4 trillion of high-yield debt matures in 2026-2027 alone, and over $5 trillion comes due through 2029, which it says will force a shift from temporary maturity extensions toward genuine restructurings, debt-for-equity swaps and changes of control, especially in commercial real estate and software 19.

Moody's default data complicates the picture further. Headline default rates are actually easing — expected default probability for U.S. listed companies fell to 7.9% in March 2026 from 9.1% a year earlier — but Moody's cautions that distressed exchanges, which let companies avoid formal bankruptcy while still imposing losses on creditors, accounted for roughly 65% of 2025 defaults 20. That means official default counts likely understate real stress, particularly in private credit, where Moody's estimates roughly 14% of borrowers don't generate enough earnings to cover current interest expense 20.

Where the reporting agrees

Across Seeking Alpha, the Financial Times, PIMCO, Reuters, CNBC and the New York Times, there is strong convergence on several points: Treasury yields have hit multi-year or multi-decade highs across major economies 13101618; the pain is concentrated at the low end of the credit-quality spectrum rather than spread evenly 141318; AI-related borrowing by hyperscalers is a genuinely new and large force in bond markets, running into the hundreds of billions of dollars in 2026 121617; and companies locked into low fixed-rate debt during 2020-2021 are shielded until those bonds mature, meaning the squeeze arrives gradually as maturities roll forward 41319. There's also agreement that Treasury Secretary Scott Bessent's buyback expansion offered only temporary relief before yields resumed climbing 8131617.

Where it doesn't

The sources diverge in emphasis and in a few specifics. PIMCO's framing is notably more reassuring than the Financial Times' or Seeking Alpha's: PIMCO explicitly argues risks are 'relatively benign' for most issuers, while the FT and Seeking Alpha lead with the alarm that junk spreads have hit levels not seen since last year's tariff shock 41315111. This is less a factual contradiction than a difference in what each outlet chooses to foreground — PIMCO is analyzing the broad universe, while the FT zooms in on the distressed tail.

There's also a live disagreement over how much AI borrowing is actually driving Treasury yields versus corporate spreads. The New York Times cites some analysts who think AI debt issuance is pulling investor money away from Treasuries and pushing government yields higher, but also cites others, including Loomis Sayles' Matt Eagan, who argue the Treasury market is too large for corporate issuance to move it much, and that the more direct impact is on corporate-bond pricing 17. IDN Financials' Reuters-sourced figure of $220 billion in hyperscaler debt issuance sits close to, but not identical with, the New York Times' figure of 'more than $200 billion,' a gap likely explained by slightly different cutoff dates or data providers rather than any real dispute 1617.

Fitch's warning that leveraged-loan and high-yield defaults could accelerate in the second half of 2026 stands in some tension with Moody's finding that default rates were easing through March 2026 — though the two are not strictly incompatible, since Moody's own analysis stresses that current calm sits atop a narrow margin of safety that a growth disappointment could quickly erode 20. That reconciliation is itself an analytical judgment rather than something the sources state outright.

The reading that holds up

Taken together, the evidence best supports a K-shaped credit market rather than a broad-based corporate credit crisis. Investment-grade and most high-yield issuers retain market access and can digest higher coupons, as PIMCO's interest-coverage data shows. But CCC-rated companies, floating-rate borrowers, and sectors like commercial real estate and legacy software face a genuine refinancing cliff as more than $5 trillion in high-yield debt comes due through 2029. The likeliest near-term consequence is not a wave of headline bankruptcies but more distressed exchanges, asset sales, deferred capital spending and quiet cost-cutting among the weakest borrowers — stress that shows up in restructuring activity and reduced investment before it shows up in default statistics.