Arini's $800 Million Fire Sale Is the First Real Clearing Price for AI-Disruption Credit Risk

About $800 million of Arini's debt changed hands in recent weeks, and the explanation was not that the firm had reached a different view on the credits. The sales were driven partly by margin calls from banks, with requirements tied to the fund's credit default swap index positions and other leveraged exposure. margin calls turned marks into sales at whatever the bid happened to be, and the resulting price becomes the new reference point for every other holder of the same paper.

Arini's losses sit in the part of the book that cannot be sold: distressed, illiquid credits in restructuring or in front of a rating action. Its cash needs sit in the part that can be sold: liquid high-yield bonds that trade every day, including Sabre paper that Bloomberg cited among the speculative-grade positions the firm has been unloading. That is why the best paper in the portfolio goes to market first. Liquid credit is being liquidated to pay for illiquid credit, which means the price Pimco and Oaktree pay sets the mark for everything else that cannot be sold.

Read the bids as a price signal rather than a credit endorsement. Pimco and Oaktree are buying the best paper from a seller with no alternatives, and every other holder of those bonds now has a reference price struck by a forced sale. The forced seller sets the price for the whole complex, and positions that cannot trade inherit that price in their marks without ever producing a trade to justify it.

In February the trade was explicit: AI-disruption fear was pushing up the cost of borrowing for software companies carrying a lot of debt, and the fear itself could cause defaults, so buy the debt everyone had marked down for disruption and earn the spread for owning it. Eight months later the same fear had flipped from the source of the discount to the source of the selling. The risk that was priced in became the reason the bonds were liquidated, and the firm that argued the fear was overdone is the one liquidating.

Credit did not break in October; it split. The option-adjusted spread on the euro high-yield index was 3.48% on Oct. 29, while the US CCC-and-lower index spread 13.25% the same day, and European CCC spreads reached their widest level since March 2020. Three weeks earlier, Fitch was still describing European high-yield spreads as low by historic standards with investor demand outstripping supply. The index kept working while the bottom of the capital structure stopped, and the cost of that gap falls hardest on borrowers whose equity story depends on an AI-disruption discount.

That gap is the AI-credit liquidity test in two numbers: a functioning market at 3.48% sitting on top of a tail where financing has effectively stopped at 13.25%. Any borrower whose capital structure now assumes an AI-disruption discount has to refinance into the second market, not the first, and the Arini sale showed what a lender chargeswhen the tail is where refinancing has stopped.

The most revealing number in the story is not a loss. Arini raised $1.5 billion for its Arini Credit strategy in October, put more than $400 million to work immediately, scheduled the rest for deployment through February, and closed its inaugural direct lending fund above $4 billion on Oct. 21, while the flagship fund was down 26% and its liquid book was being sold. The obvious response to a public-credit liquidity event is to raise more money for credit that does not mark daily, and that migration moves the same risk somewhere it is harder to exit and easier to forget.

Blue Owl Capital Corporation is the cleanest listed marker for what that migration does to marks. Net assets were $7.03 billion, or $14.26 a share, at June 30, 2026, against $7.40 billion, or $14.81 a share, at Dec. 31, 2025; borrowings stood at $6.1 billion against that $7.0 billion of net assets, and non-accruals were 2.8% of the portfolio at cost and 0.8% at fair value across 229 companies and $15.0 billion of fair value. On that arithmetic, a further 8% markdown on the portfolio would erase roughly $1.2 billion, about 17% of net assets. Private marks import the public mark with a lag, and the lag is where the next repricing lands.

The AI credit complex is why to care about the tail. Goldman Sachs reckons roughly a third of AI capital spending in 2026 will be debt financed, and Morgan Stanley projects private credit will supply a further $800 billion of data-centre financing over the next two years. That is a large pipeline of new paper, and the Arini sale has just shown what the marginal buyer of risk is paying attention to. AI borrowers now compete with leveraged holders of the same credit for the balance sheet that prices it.

The arithmetic matters: CoreWeave's trailing interest expense of $1.87 billion consumes 47.5% of trailing EBITDA of $3.94 billion, against revenue backlog of $99.4 billion at March 31, 2026 and revenue compounding from $1.36 billion to $2.58 billion between the September 2025 and June 2026 quarters. The asset base is not the problem; the funding mix is. A wider distressed tail raises the coupon on every incremental dollar of AI data-centre debt, and the first place that shows up is the next tranche priced in the fourth quarter, with third-quarter results due Nov. 10, 2026.

One listed balance sheet shows how thin the equity cushion at the bottom of the market has become. Sabre ended June with $4.32 billion of debt, $676 million of cash and negative book equity of $1.07 billion; trailing interest expense of $207 million now exceeds trailing EBITDA of $174 million, and trailing operating cash flow of $75 million does not cover the interest bill. Its bonds are the liquid example of what is being unloaded. The equity in these issuers is a call option on a refinancing that just got more expensive.