Tightening credit hitting AI boom amid a “bloodbath” for junk bonds

5 November 2026facebook iconThe storms which ripped through Wall Street last week, leading to $900 billion being wiped off the market value of AI and high-tech companies, have passed—for the moment.

But the forces which gave rise to them continue to intensify, preparing the conditions for another eruption on a bigger scale. Reports and analysis across the board confirm this fact of financial life.

The major AI firms continue to spend money on the AI buildout and the development of computing power, running up larger debts even as financial conditions become more difficult and risky.

But they cannot call a halt, because to do so would mean falling behind in the race to grab a portion of the profit bonanza they have gambled on securing, even as analysis shows that the revenue generated by AI will fall short of that needed to cover all the capital spending and secure a sufficient return.

Diseases of the human body often manifest themselves at its extremities before striking at the essential organs. There is a parallel in the financial system.

Earlier this week the Financial Times (FT) pointed to what has been characterized as a “bloodbath” in the so-called junk bond market, particularly in Europe.

“A prolonged rout in the riskiest parts of the global markets for junk debt is sending shockwaves through credit markets, driving some of its most active investors to steep losses as investors lose confidence in the world’s most vulnerable corporate borrowers,” it reported.

The yield or interest rate on CCC-rated bonds, the lowest rung of junk debt, has risen to as high as 21 percent in Europe and 18 percent in the US, up from around 15 percent at the start of the year.

According to an unnamed portfolio manager at a large US hedge fund cited by the FT: “It’s a complete bloodbath. Everyone is scared of being in the next blow up … being flat is the new up.”

The concerns go beyond the junk bond market to the development of AI which has powered Wall Street over at least the past three years.

Last month, the $80 billion US hedge fund Elliott Investment Management warned in an investor letter, cited by the Wall Street Journal, that markets were coming to resemble the Roaring ’20s and that “at present, virtually every alarm bell warning of a substantial bear market ahead is ringing.”

The letter warned that circular financial arrangements between the companies that develop AI models and those that build the infrastructure, the increase in debt-financed capital spending, the growth of an anti-AI movement and diminished AI expectations “could be historic, or at least as severe as any bear market in the last few decades.”

“So many factors have lined up to cause this great boom that it might take disruption in only one or two of them to shift this momentum into reverse, leading to a cascading, self-reinforcing unwind of this bull market and an abrupt change from FOMO [fear of missing out] to FOGO [fear of getting obliterated],” the letter said.

Chief among these factors is the free cash flow position of the hyperscalers financing the technology. The free cash flow is what remains on hand after paying the costs of the business and capital expenditure. Initially this remained positive as companies such as Amazon, Google and others were able to finance their capital spending out of their cash holdings.

That has now ended. It was one of the issues which led to the fall in the shares of Alphabet (parent of Google) last week, following the revelation that its free cash flow had come in at minus $9.1 billion in the September quarter compared to minus $5.9 billion for the June quarter.

All told, the four largest hyperscalers—Alphabet, Microsoft, Meta and Amazon—had negative cash flows of $45 billion in the quarter compared to positive cash flow of $36 billion a year before.

This means AI development must increasingly be funded by debt or the issuing of new shares. But a turn to debt financing takes place in very different conditions from those which prevailed for almost a decade and a half after the crash of 2008, when interest rates were at historic lows.

The raising of money now takes place in competition with governments in the US and around the world that must issue new debt just to pay the interest bill on their previous borrowing, which is assuming an ever-increasing portion of their budgets.

As Ruchir Sharma, the chair of Rockefeller International, explained in a comment in the FT earlier this week: “The pool of available capital is more limited than it was in the booms of the late 1990s and 2000s, because this time companies must compete with governments for money.”

“This sets the stage for a rapid liquidity squeeze,” he continued, noting that AI-related debt issuance is likely to exceed $600 billion in 2026, a 10-fold increase from 2024.

How does this AI-driven credit squeeze compare historically to the 2008 financial crash, in terms of the mechanisms that triggered the crisis?What happened to workers' jobs and living standards after the dot-com bubble burst in 2000, and does that offer a warning for what could follow an AI bust?Ask more questions at SocialismAI.comA “credit winter” has not yet come for the major AI companies, but there are indications it is on its way. On Monday, Oracle sold bonds with a coupon rate—the interest rate the borrower agrees to pay on the initial par value of the bond—of 7 to more than 8.5 percent compared to the 5 to 7 percent it had to pay in February.

The cost of insuring major AI companies against default, via credit default swaps, is rising. The cost to insure Alphabet’s debt is just below 1 percent of the principal, on a par with that of Mexico, the debt of which is at the lowest level of investment grade.

An analysis on Bloomberg said the credit market was delivering early warnings it was approaching the limit of debt it could absorb.

It cited remarks by Elon Musk that hundreds of so-called neocloud firms which service AI could go under because they could not refinance, as well as cautions issued by the hedge fund Citadel that even the biggest firms could be ensnared in debt problems.

A report by the research firm Gartner, which has already warned that AI revenue could fall trillions of dollars short of what is needed to pay for the capital spending and ensure a profit, said that 80 percent of neocloud firms could go out of business.

The Bloomberg analysis said that while there were no predictions of an immediate credit crunch, in many ways it was beside the point.

“The concern is that, with trillions more in debt needed over the coming years, there appears to be a growing disconnect between what Silicon Valley’s top executives believe can be raised, what financiers say they can deliver, and what money managers are willing to take down.”

The pending initial public offering (IPO) by Anthropic, scheduled for later this month for which the prospectus has been issued, has been characterized as “bellwether” for the market. A New York Times report described Anthropic as a “tent pole of the AI industry”—the implication being that if the IPO fails or falls short the tent could collapse.

Doubts have already started to emerge with the private capital giant Blackstone, a major anchor for the IPO, telling investors Anthropic may not receive investment grade status on going public, contrary to what was said previously. The rating is important because if it receives anything less than investment grade, Anthropic will be cut off from important sections of capital because many pension and insurance funds would not be allowed to make large investments in it.

If it succeeds and achieves a market value of $1.75 trillion or above, this could set off a new surge in the AI bubble, amid accompanying hype. But the effect would only be to deepen the contradictions in the AI boom now rapidly coming to the surface.