This has been so long in the making. Gradually… then suddenly… Repeating a comment I’ve stated over the years: I hope my analysis is wrong. I fear the greatest Bubble in human history has finally begun to falter. An especially disconcerting period has commenced.

I posited last week that the unfolding crisis was systemic, with the U.S. and the AI arms race at the epicenter of an unprecedented global Bubble.

The Semiconductor (SOX) Index was hammered 11.4% this week, the largest weekly loss since “liberation day.” Supporting the bursting Bubble thesis, the semi equipment stocks were under notable pressure. Applied Materials lost 19.3%, Lam Research 17.2%, and KLA 15.3%. Coherent dropped 25.5%, Lattice Semiconductor 17.8%, and Marvell Technology 17.2%. Nvidia slumped 6.1%, down a quick 12% in eight sessions. Broadcom fell 9.1% to a six-month low. Micron sank 14.6%, Intel 14.2%, and AMD 13.1%.

South Korea’s KOSPI Index was hammered 14.5% this week, with a seven-session drop of 18.3%. The index ended the week down almost 40% from its June 19th all-time high. Samsung sank 17.7% this week, with SK Hynix down 24.3%. From June highs, Samsung has slumped 43% and SK Hynix 57%.

Trading this week to a one-year low, the Shanghai Composite fell 2.6%. China’s growth-oriented ChiNext Index sank 8.9%, with losses from June highs at 37%.

Japan’s Nikkei Index fell 5.4%. AI heavyweight SoftBank sank 18.2%, with losses from June highs now at 45%. Taiwan’s TAIEX Index dropped 4.5% this week, with Taiwan Semiconductor losing 5.1% (down 6% from October 6th record high).

It has become a bad habit. Another week and yet more “holy crap” moments.

Ten-year Treasury yields surged 12 bps to 5.33% on Wednesday’s disappointing CPI report – a level that would have been the highest close since 2007. Importantly, inflation and Treasury yields were not at the top of market concerns.

October 15 – Financial Times (Ramsay Hodgson, Michelle Chan, Euan Healy and Ian Smith): “Investors are losing confidence that revenues from the AI boom will be sufficient to support the vast debt piles financing data centres, pushing bonds backing some of the largest US infrastructure projects into increasingly stressed territory. Prices of bonds issued to fund prominent data centres where Meta, Nvidia, Oracle and others will become tenants — such as ‘Hyperion’ and ‘Project Excalibur’ — have fallen well below par in recent weeks. Investors are also dumping debt issued by so-called neocloud companies — of which the best known is Nvidia-backed CoreWeave — which have been the most aggressive in how they finance their AI build-out.”

CoreWeave bond yields surged 135 bps intraday to a Wednesday high of 15.82% (up 220bps from Monday’s close and 450 bps over two months). Oracle (’36) yields jumped 18 bps to a record 7.78%, while Meta (’36) yields spiked 15 bps to a record 6.35% (Oracle yields ended June 29th at 6.04% and Meta yields at 5.27%). High yield corporate bond CDS spiked 10 bps intraday Wednesday, trading above 350 bps for the first time since April. Markets for key AI-related debt were dislocating.

Curiously, Treasury bonds rallied sharply post-CPI spike, with yields reversing 14 bps lower to end the session at 5.19%. Safe haven buying gained momentum Thursday, with 10-year yields dropping 13 bps to 5.06%. Surging Treasury prices are a problem for key levered strategies, such as “carry trades,” where Treasuries are the funding source for higher-yielding corporate and other debt securities. At Wednesday’s close, the high yield spread to Treasuries had widened 16 bps in three sessions to 3.24 percentage points – the widest spread since the March war spike. Deleveraging.

October 16 – Axios (Madison Mills): “Wall Street appears to be hitting its limit on AI, and financiers are starting to say the quiet part out loud: All of this AI spending doesn’t look like it can pay off, at least not for everyone. The gap between what it costs to build AI infrastructure and what companies are actually paying to use it is widening. Wall Street’s anxiety is surging as the debt piling up to fund data centers and chips looks increasingly disconnected from economic reality. A massive gap is opening up between AI infrastructure spending and expected returns. While AI providers are projected to spend $7 trillion on infrastructure through 2030, research from Gartner forecasts they will generate no more than $3.6 trillion in related revenue over that same period. More than $4 trillion in AI infrastructure spending could be at risk of write-downs by the end of 2032… Gartner also predicts that more than 80% of specialized AI cloud providers, known as neoclouds, will cease operating independently by 2030.”

At Thursday morning’s lows, CoreWeave’s stock was down 27% w-t-d. A 17% intraday rally saw the stock end Thursday’s session with a 6% gain.

October 15 – Reuters (Aditya Soni, Anhata Rooprai and Laura Matthews): “Shares of Nvidia-backed CoreWeave steadied on Thursday after a sharp slump, as investors worried about the debt financing behind the artificial intelligence build-out. The loss-making cloud company has about $50 billion in debt and has become a focus of concerns that the data-center boom is being fueled by credit, straining newer players and pressuring cash flows even at giants such as Alphabet and Meta Platforms… Nvidia CEO Jensen Huang, whose company is CoreWeave’s second-largest investor with about a 10% stake, played down the fears. ‘CoreWeave’s financial strength is improving every day,’ Huang told reporters…, adding that he does not see a need for Nvidia to backstop the company. ‘I’m optimistic. Their cost of capital is actually decreasing.’”

“Decreasing” “cost of capital” is an interesting take. CoreWeave’s bond yields, 9.625% at June 18th issuance, ended the week at 15%, while the stock closed Friday down 14% y-t-d.

Jensen Huang’s steadfast optimism was widely shared. Despite intensifying debt market problems, the VIX (equities) Index closed Thursday’s trading at a relatively serene 16.46. That susceptible dam gave way Friday.

“CoreWeave’s Big Dream Crashes Into Reality as Yields Top 15%.” “Why CoreWeave’s Woes Matter to the Market.” “CoreWeave Bond Rout Tests Market’s Math on AI Buildout.” “AI Investors Have Woken up to the Bond Market’s Warning Signs.” “CoreWeave AI Selloff Reveals Fault Lines Across High Yield Tech.” “Nvidia Partner GMI Cloud Calls CoreWeave’s 15% Yield ‘Insane.’”

CoreWeave’s stock was slammed 12% in ominous Friday trading. Fellow “neocloud” company Nebius saw its stock sink 12.2%. Palo Alto Networks dropped 7.6% and Palantir fell 7.2%. After trading to a record high on Tuesday, the MAG7 Index reversed 5% lower, with the index down 1.9% in Friday trading. Friday losses included Meta 3.7%, Alphabet 2.5%, Tesla 2.1%, Microsoft 1.9%, and Amazon 1.9%. Market dynamics are reminiscent of the final speculative run (in the face of deteriorating industry fundamentals) for the big tech stocks that reversed sharply from all-time highs in late-March 2000.

The unfolding debt crisis gathered momentum Friday. CoreWeave CDS surged another 42 bps to a record 1,128, spiking an alarming 226 bps this week. High yield CDS jumped 10 bps Friday to 357 bps – the highest close since April 7th. Spreads for technology bonds within the Bloomberg US Corporate High Yield Index widened a notable 19 bps Friday. Investment-grade CDS gained two to 62.2 bps – the high since March 31st. Nvidia CDS gained another two (10 for the week) to a record 96 bps, with SoftBank CDS trading to a record 428 bps Thursday (closed the week up 22 to 421 bps).

October 13 – Reuters (Anton Bridge): “SoftBank’s promise to pay its highest-ever yields in a record $11 billion junk-bond offering highlights the pressure on CEO Masayoshi Son to secure returns on his all-in bet on AI amid concerns about possible overheating in the market. The SoftBank debt issuance last month added to a sharp surge in corporate borrowing this year, including major bond sales from Amazon.com, Alphabet and Tencent. The scale of Son’s investments, which include backing totalling $64.6 billion for ChatGPT maker OpenAI, in particular, has made the technology-and-investing conglomerate a bellwether for AI sentiment. Those fresh investments are now coming under sharper scrutiny by analysts and investors amid worries about the degree of concentration in SoftBank’s portfolio and whether a payoff from the expenditures by AI companies will come soon enough.”

SoftBank (’34) yields surged a notable 33 bps in Friday trading to a record 9.95%, with yields spiking 49 bps on the week. Skydance’s high-yield (’36) bond yields jumped another 42 bps this week to 10.58% (77bps in two weeks). Some large and levered companies are experiencing alarming bond yield spikes.

I’ll take special note of widening high yield spreads (to Treasuries). Spreads surged another 10 bps Friday to 331 bps, boosting the week’s gain to 23 bps, and the four-week spike to a noteworthy 64 bps. High yield spreads closed the week within four bps of the war-induced March spike high.

Over the years, I’ve posited analysis of a particularly problematic market dynamic for the leveraged speculating community: this would be the combination of rising market yields, widening spreads, and surging CDS (market “insurance”) prices. This dynamic took decisive hold this week, portending accelerating de-risking/deleveraging, waning market liquidity, and worsening debt issues.

Leveraged Loan (LSTA index) prices dropped 24 cents this week to 94.73 – the low since April 13th. Indicative of the rapidly deteriorating backdrop for leveraged finance, this index is down almost a full point in a month – to within 14 cents of the low back to July 2023.

It’s also worth noting this week’s 10.3% slump in the Goldman Sachs Most Short Index – the steepest weekly decline since “liberation day.”

I closely monitor this index’s performance as a valuable financial conditions indicator. The index is typically comprised of levered companies of questionable underlying fundamentals (the reason behind outsized, chiefly hedge fund short positions). This index will generally outperform in rising markets, where loose conditions mask fundamental company issues, while supporting short squeezes and general risk-taking.

But this index invariably suffers when risk aversion and deleveraging begin to take hold. Hedge funds often move quickly to build short exposure in suspect companies as a market hedging mechanism. Meanwhile, a tightening of market liquidity and Credit conditions poses major risk to companies with excessive debt, negative cash flows, and over-ambitious/suspect business models (not to mention fraud and chicanery).

This week’s short index bludgeoning corroborated the thesis of an unfolding consequential tightening of market liquidity and Credit conditions. The 3% slump in the small cap Russell 2000 provided additional confirmation, as did the 3% drop in the Broker/Dealer Index.

Reality seemed to take hold as Friday’s session unfolded. The VIX Index traded below 16 in early trading, only to spike above 22 (four-month high) later in the session (closed the week at 19.98).

I infer a low VIX in the face of mounting debt market issues as stock market confidence that a big sell-off will be avoided ahead of the midterms. We’re down to 18 days, and by Friday afternoon market sentiment had shifted to fear of various risks that might erupt at any point.

October 16 – Reuters (Summer Zhen and Anirban Sen): “Global hedge funds suffered steep losses this week, battered by a sharp unwind of a swathe of AI bets, with Asia-focused funds hit the hardest… Global fundamental long-short funds, or stock pickers, lost money every single day this week, the Goldman Sachs prime brokerage unit said…, with losses for the week to October 15 coming to 2.9%. For the year to date, they have gains of 14.3%… On Thursday, hedge funds across strategies retreated sharply… ‘What we saw was less a fundamental rejection of AI and more a classic, high-velocity unwind of a crowded trade meeting a macro reality check,’ said Bruno Schneller, managing partner at Erlen Capital Management, which invests in hedge funds… ‘The interesting thing about this selloff is that it exposes just how much of the hedge fund industry has effectively been trading the same underlying theme, even when portfolios look very different on paper,’ said Schneller.”

Especially after several years of historic global speculation, liquidity overabundance, and loose financial conditions, the proliferation of crowded trades and leverage becomes a serious issue. Hedge funds “retreated sharply” Thursday. Friday was much worse, which sets the stage for heightened de-risking/deleveraging next week. The crowding problem goes much beyond the leveraged speculating community.

October 13 – Bloomberg (Julien Ponthus): “Global investors chasing the AI-powered equity rally are the most exposed to technology stocks since the rebound from the Covid pandemic six years ago, according to a poll by Bank of America Corp. The proportion of fund managers who are overweight the tech sector has risen to 37% from 30% last month, the highest since August 2020… About 45% of respondents agreed that there will be an AI bubble in 2027, with 42% saying no.”

Fragility has reached the point where the administration will surely seek to reverse market sentiment. And, of course, forceful short squeezes and the unwind of bearish hedges could materialize anytime. The President’s claim that Iran is ready for a deal would be skeptically received in the marketplace.

Friday’s $3.44 surge in WTI crude pushed the week’s advance to 10.4% – with Friday’s $101.41 close the highest since May. Iran expanded tanker attacks this week, while the Houthis continued missile strikes on Saudi infrastructure. There is also the possibility that energy markets are bracing for a return to major military conflict. The possibility of the nightmare scenario, with major additional damage to Gulf energy infrastructure in a backdrop of severely depleted global crude and refined product inventories, supports elevated prices.

“Wall Street Anxiety Spreads as AI Rout Meets $110 Oil Shock.” “AI and Chip Stocks Shed Over $1 trillion in a Week.” “New Cracks in AI Trade Push Wall Street to Worst Week Since June.” “The Debt-Fueled AI Boom is Cracking.”

October 16 – Bloomberg (Ian King): “Nvidia Corp. CEO Jensen Huang took a sanguine view toward the latest AI stock-market rout, saying the benefits of the technology will soon outweigh the concerns of investors. ‘All of this is going to pass, and the reason for that is very simple,’ Huang told Bloomberg’s Emily Chang… ‘It’s going to be transformative. It’s going to impact in a very positive way every single industry and every single country. There’s no question it will continue, and we’re at the beginning of that.”

The analytical backdrop is anything but “very simple.” AI is “transformative” and will undoubtedly continue. But this unfolding debt crisis will not pass before inflicting tremendous damage. When it comes to AI debt, the bloom is off the rose.

October 16 – Financial Times (Michelle Chan and George Hammond): “Anthropic is offering Blackstone a high yield of more than 10% to finance its latest chip deal, highlighting the start-up’s willingness to stomach steep borrowing costs to finance expensive AI infrastructure. The $60bn debt package, dubbed ‘Project North Star’, would help fund Anthropic’s procurement of about 1.5 gigawatts of custom semiconductors developed by Google… The private capital giant was offered a coupon of 10% and an even higher total return after accounting for new-issue discounts and other fees, the people said. A borrower with a similar credit profile typically pays a yield of about 7% in the public bond market… Just four months ago Anthropic sold about $4.4bn of debt in a similar deal known as ‘Project Big Sky’ at an interest rate of 8.5%…”

I don’t see how Trillions of data center and AI investment work at double-digit market yields. This has been historic late-cycle crazy excess that is now quickly returning to haunt system stability. The bottom line is that debt to fund a historic global AI arms race is extraordinarily risky. Costs have skyrocketed, with revenue prospects turning murkier by the week. Previously unappreciated costs and risks are moving into focus. The system will now be burdened with hundreds of billions of additional “junk” Credit, while hundreds of billions of additional debt will transform the investment-grade Credit universe. I wouldn’t count on market risk aversion and instability “passing” anytime soon.

October 15 – Wall Street Journal (Jean Eaglesham): “A string of artificial-intelligence safety incidents is exposing a gaping hole for the fast-growing industry: the lack of available insurance. Rogue agents have hacked into websites. AI chatbots have been accused of encouraging suicides and murders. An agent submitted fake unsolved murder tips. The latest episodes add to existing lawsuits AI companies face alleging chatbot harm and copyright violations when models are trained. The seemingly limitless possibilities of AI to change the world also include a wide-ranging capacity for harm that could leave companies that produce or use AI legally vulnerable. The biggest AI companies are racking up potential legal liabilities running into the tens—if not hundreds—of billions of dollars…”

WSJ: “AI Runs Into Its Next Big Obstacle: Insurance.” One is hard pressed to view AI-related risks as even insurable losses. The astronomical costs involved if something really runs amok are impossible to quantify. How would insurance actuaries quantify and price for such risks? And on the subject of insurance companies…

October 13 – Financial Times (Ryan McMorrow): “CoreWeave is chasing insurers’ ‘unlimited’ pools of capital as the Nvidia-backed AI cloud provider seeks cheaper ways to fund the tens of billions of dollars it is spending on chips and other AI infrastructure. The push comes as the AI boom increasingly leans on debt deals to build data centres and buy the Nvidia chips needed to power AI models. That has made life insurers — and the trillions of dollars they manage — the biggest prize for borrowers seeking cheaper funding. Brannin McBee, a co-founder who leads CoreWeave’s financing efforts, said winning investment-grade ratings for more of its chip-backed debt, and eventually the company itself, would allow it to borrow from insurers whose capital ‘is unlimited. Like, that’s all the money in the world that sits in the insurance tranche.’”

The quiet part out loud: “borrow from insurers whose capital ‘is unlimited. Like, that’s all the money in the world…” It sure appears the hundreds of billions worth of unsuspecting annuity buyers have been this cycle’s “dumb money.” If private Credit problems haven’t already awoken the insurance industry, I’ll assume AI debt problems now sound the alarms.

The market still prices one Fed rate increase by yearend – and a second by the March 17th FOMC meeting. I wouldn’t be surprised by one and done – with the one being last month’s rate increase.

If de-risking/deleveraging unfolds as I expect, market attention will shift to the Fed’s balance sheet and need for central bank liquidity support. This is problematic subject matter. Especially after this week’s 3.7% y-o-y September CPI reading, the last thing the Fed is contemplating right now is a big surge in assets (liquidity injections). But if the levered players are turning nervous, they don’t want to ponder a Federal Reserve reacting slowly and conservatively to market liquidity issues.

And if we jump forward and war-game crisis dynamics, there are some uncomfortable aspects to the current backdrop. For one, the world has so much more debt than 2008. The amount of global speculative leverage has skyrocketed. Internationally, economies are much more imbalanced and structurally maladjusted. And it’s not reasonable to assume that a concerted crisis response from the world’s leading governments and central banks would not be forthcoming in a timely manner.

Today’s vulnerabilities are indeed frightening, including two wars, both with the potential to spiral out of control. Meanwhile, President Trump is more erratic and volatile than ever. The odds don’t seem good, but we could enjoy $5,000 post-midterm checks. Or we might be 18 days from mayhem. I don’t discount the possibility of a “scorched earth” turn in the event of a big blue wave. And Tehran, transfixed by tottering global markets, must feel more emboldened by the week. Global market fragilities surely inspire an even harder line.

If their goal is to punish the Trump administration and the U.S. to ward off future military hostilities, the Iranian hardliners have an opening. To find ourselves today with myriad serious vulnerabilities is incredibly frustrating. And we have reason to worry about how our mercurial President will operate in an environment increasingly disorienting and combative on multiple fronts.

October 16 – Wall Street Journal (Peter Landers): “Higher interest rates and persistent inflation in the U.S. and elsewhere have upturned the way Moody’s assesses risk, according to the ratings company’s chief credit officer. ‘We believe we are in a new macro regime,’ said Atsi Sheth, who leads the development of Moody’s global credit views. Leverage—the extent of a borrower’s debts relative to assets—matters more now because interest costs are rising, Sheth said… ‘The more leveraged you are, the more higher rates will likely hurt you,’ Sheth said. That contrasts with what she described as a ‘fairly benign environment’ in the 2010s, when inflation and interest rates were generally low.”

Original Post 16 October 2026

Categories: Doug Noland