This is even better than my sub-prime gig!
By Scott Ortkiese | June 2, 2026 | so@throughlinesynthesis.com
The BlackRock story is not a footnote. It is a warning flare over a $1.8 trillion market that has been operating with the discipline of a casino back office.
Phil Tseng, CEO of BlackRock TCP Capital Corp., is out. His departure follows months of losses on soured loans, two separate net asset value markdowns totaling roughly 24 percent in five months, and a federal probe by the Manhattan U.S. Attorney’s Office into whether the fund was deliberately marking its loan book above true value to protect fee income. The Southern District is run by Jay Clayton, who has already gone on record saying that mismarking assets to collect fees “has always been a no-no.” That is not background noise. That is a former SEC chairman signaling that the Justice Department is now looking at the plumbing of the private credit industry the same way it eventually looked at CDO valuations in 2007.
Here is the problem. Business development companies like BlackRock TCP Capital are not obscure boutique vehicles. They are publicly traded wrappers that pool private credit loans, many of them made to AI infrastructure plays, data center developers, leveraged buyout targets and tech startups burning cash in the hope that the AI capex cycle never ends. The loans inside them do not trade on any active market. Valuations are set by the fund managers themselves on a quarterly basis. Investors who put money in are often locked out, told their redemption requests cannot be honored, and left reading quarterly marks that may or may not reflect what the loans are actually worth. That is not a bug in the design. That is the design, and it functioned as long as everyone agreed not to ask hard questions.
PIMCO has been asking hard questions. Christian Stracke, president of the $2.3 trillion firm, said flatly in March that the private credit market is experiencing “a crisis of really bad underwriting” and that it “should eventually face a full-blown default cycle.” PIMCO’s strategist Lotfi Karoui has since published a warning about a “confidence gap” between managers whose loan quality is sound and those whose books are driven by optimistic assumptions rather than facts. The industry saw its first net outflows in Q1 2026, reversing what had been robust inflows. JPMorgan has already started restricting certain lending to private credit funds after marking down loans on their own books. Cliffwater’s $33 billion fund is facing redemption pressure. BlackRock’s HPS Corporate Lending Fund capped withdrawals at 5 percent after investors asked for nearly double that.
This is the same movie from 2007 with a different cast. Then it was mortgage-backed securities bundled by ratings agencies who vouched for their quality while the underlying loans rotted. Now it is private credit BDCs bundling loans to AI data centers and leveraged companies, marked by fund managers themselves, with investors locked in and federal prosecutors starting to ask whether the marks were ever honest. The opacity is worse this time because at least mortgage loans had a secondary market. Private credit loans have quarterly manager estimates and that is it.
The AI capex question sits underneath all of this like a fault line. Roughly 34 percent of global fund managers surveyed by Bank of America identified AI hyperscaler capital spending as a likely source of the next systemic credit event. That number had doubled since April. If AI infrastructure delivers on its promises, some of these loans survive. If it does not, if the data centers get half-built, the startups fail to monetize and the capex cycle turns off because the Fed decides it needs to defend price stability over AI dreams, then the private credit market does not “grind through stagflation.” It blows up.
The Ford rehiring story from the transcript makes the point in the most human possible terms. Ford laid off experienced engineers, replaced them with AI systems, and then had to rehire the same people because the AI could not handle quality problems that humans solve without thinking. If that pattern repeats across industries, the productivity miracle that justifies the AI capex never materializes. The borrowers stop servicing the loans. The BDCs mark down their books again. And this time, unlike January’s 19 percent NAV cut that shocked the market, the numbers will not be a surprise to anyone paying attention because PIMCO, the DOJ, and the departing CEO of BlackRock TCP Capital have already told the story out loud.
The Fed is the final variable. Every overleveraged player in the AI credit complex needs rates to stay down or go lower. Musk’s SpaceX reportedly needs cheap money to service old debt with new debt. Data center developers need cheap money to keep building before the business case is proven. BDC managers need cheap money to stop their weakest borrowers from defaulting and triggering marks they cannot hide. But the Fed has signaled it is in the price stability business first. When those two needs collide, the people holding the unliquidated, opaquely marked, quarterly-reported private credit paper will discover that illiquidity is not a feature. It is the trap.
The stagflation crowd says this is all manageable. A reckoning in private credit, some defaults, some fund blowups, a painful but contained adjustment. That reading requires you to believe the $1.8 trillion private credit market is sealed off from the Korean chip crash, the U.S. AI equity unwind, the Fed’s rate posture, and the capital rotation away from Western assets toward China, Russia and the energy sovereigns of the Middle East. None of those things are sealed off from each other. They are the same story told from different vantage points, and the BlackRock probe is simply the moment the credit layer of that story stepped into public view.
Related reading
- Europe’s Private Credit Market Is Unsupervised and Untested
- Генерал «Предай-США» (Betray-U.S.), он же Дэвид Петреус, идеально иллюстрирует оксюморон: военная разведка
- General Betray-Us, aka David Petraeus, Perfectly Illustrates the Oxymoron: Military Intelligence
- AI Revenue Was Always the Lie, and the Time Bomb Is T-Minus 18 Months and Counting, Not Four Years
