“Unintended Consequences is full of substance, it is one of the must-read books of the year, and once I finish it I will be giving it a second read through right away.” - Tyler Cowen, Professor, George Mason University
“Unintended Consequences provides a provocative interpretation of the causes of the global financial crisis and the policies needed to return to rapid growth. Whether you agree or not, this analysis is well worth reading.” - Nouriel Roubini, New York University; Chairman, Roubini Global Economics
“…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
“…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
“…reminds us that inequality sends a signal of what society lacks most, in America’s case, entrepreneurship and risk taking.” - Lawrence Lindsey, CEO, The Lindsey Group, former Director of the National Economic Council
“A full-throated defense of economic dynamism.” - The Wall Street Journal
“Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
“…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
“There are an amazing number of good ideas and interesting points made in Unintended Consequences. The thinking underlying it, and the obvious depth of understanding of the author, are very impressive.” - Steven Levitt, coauthor of Freakonomics; 2004 John Bates Clark Medal
“Unintended Consequences represents the most cogent and persuasive analysis of the Financial Crisis to date.” - Andrei Shleifer, 1999 John Bates Clark Medal Winner
“…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
“…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
AI Summary.An "AI universe" index of 42 stocks groups companies into three categories: 28 direct AI firms including NVIDIA, Microsoft, Alphabet, Amazon, Meta and Broadcom; 8 AI-linked utilities such as Vistra, NextEra and Constellation; and 6 capital equipment makers including Eaton, Quanta and GE Vernova.
Edward ConardMacro Roundup
Date Posted:
As the probability of a Democrat-controlled Senate rose from 50% in online betting in early September to 100% after the election, the S&P 500 fell 4.7% lead by an 8.9% drop in AI-related stocks and a 1.1% decline in the rest of the market.
How is the AI universe index of stocks composed?
Core argument: Michael Cembalest's "AI universe" comprises 42 stocks across three categories: 28 direct AI names, 8 AI utilities, and 6 AI capital equipment companies.
As the probability of a Democrat-controlled Senate rose from 50% in online betting in early September to 100% after the election, the S&P 500 fell 4.7% lead by an 8.9% drop in AI-related stocks and a 1.1% decline in the rest of the market.
Takeaways by Macro Roundup® AI
Michael Cembalest's "AI universe" comprises 42 stocks across three categories: 28 direct AI names, 8 AI utilities, and 6 AI capital equipment companies.
The 28 direct AI stocks in Cembalest's universe include NVIDIA, Microsoft, Apple, Alphabet, Amazon, Meta, Broadcom, Tesla, Oracle, Palantir, AMD, Salesforce, IBM, Intel, Micron, Dell and Super Micro Computer.
Cembalest's AI universe includes 8 utilities, among them NRG, Vistra, NextEra and Constellation, and 6 capital equipment makers, among them Eaton, Trane, Quanta and GE Vernova.
Related Articles:
The Stock Market’s Read On The Midterms, In Its Winners And Losers— Stocks rose broadly after a sweeping Democratic win in midterm elections, with health insurers tied to the Affordable Care Act and solar companies gaining while data-center firms and companies linked to Trump and Musk fell. The move reflects the increment beyond already-priced-in Democratic victories, from greater certainty and a larger-than-expected margin.
Califoreboding— A basket of 42 AI stocks accounts for 60%-80% of all S&P 500 earnings growth, price returns and capital spending, while 60% of S&P 500 stocks sit 20% or more below their highs and only 30% outperform the index. AI-related construction spending has risen 38%, the only positive trend as manufacturing, commercial real estate and residential construction decline.
One Hundred Years in the U.S. Stock Markets— Over a century, U.S. publicly traded stocks generated roughly $91tn in shareholder wealth enhancement, with returns strongly positively skewed and dominated by a relatively small number of high-performing firms. Because most possible outcomes fall below the mean in a skewed distribution, the authors argue financial planning should focus on median outcomes or the full distribution.
AI Summary.Financial firms, companies and governments raised about $6.61tn through broadly syndicated investment-grade and junk-bond sales, up 15% year-over-year to a record. Cloud-computing giants and data-center developers sold almost $368bn of debt, more than 6x the prior year, accounting for roughly a third of the increase in borrowing.
Gerson Freitas Jr. and Ronan MartinBloomberg
Date Posted:
Bloomberg data shows that global investment-grade and high-yield bond sales, counting only governments’ syndicated deals and not their auctions, have hit a record ~$6.61T year-to-date, up 15%. AI-related issuers account for about a third of the increase.
Is AI borrowing driving global bond sales to a record?
Core argument: Global broadly syndicated investment-grade and junk-bond sales reached about $6.61 trillion so far this year, up 15% from a year earlier and eclipsing 2025’s record with two months still to go, according to data compiled by Bloomberg.
Financial firms, companies and governments have raised a combined ~$6.61 trillion through broadly syndicated investment-grade and junk-bond sales so far this year, up 15% from the same period a year earlier and eclipsing 2025’s record with two months still to go, according to data compiled by Bloomberg. Cloud-computing giants and data-center developers alone sold almost $368 billion in debt, more than six times as much as a year ago, the data shows. The group has accounted for about a third of the increase in borrowing so far this year, outpacing every other non-financial sector and adding twice as much incremental debt as governments.
Takeaways by Macro Roundup® AI
Global broadly syndicated investment-grade and junk-bond sales reached about $6.61 trillion so far this year, up 15% from a year earlier and eclipsing 2025’s record with two months still to go, according to data compiled by Bloomberg.
Cloud-computing giants and data-center developers sold almost $368 billion in debt, more than six times the amount a year earlier, according to data compiled by Bloomberg.
Cloud-computing and data-center borrowers accounted for about a third of this year’s increase in borrowing, outpacing every other non-financial sector and adding twice as much incremental debt as governments.
Related Articles:
Wall Street’s AI Debt Machine Is Starting to Choke on Supply— At least 18 data-center financings in the junk-bond and leveraged loan markets carry double-digit yields, and a $5bn syndicated loan for neocloud Volta Infrastructure required steeper discounts and higher interest, pushing its yield above 11%. The average yield on investment-grade technology bonds has risen to about 6%, roughly 1.4 percentage points higher than a year earlier.
The Post‑Pandemic Global R*— Using an econometric model extracting the common trend in real interest rates across advanced economies, the authors estimate that the global neutral real interest rate fell about 3.5 percentage points from 1990 to 2019, then rose about 1 percentage point after COVID, leaving global and U.S. estimates near 0.5%.
AI Set to be Largest CapEx Cycle Ever … and Soon Majority Externally Financed— Hyperscaler and data-center investment is tracking toward $1.5tn annually by 2030, which in real terms would exceed the 2005-06 housing boom, the 1990s fiber peak, and the 1882 railroad peak combined. External debt and equity are projected to fund close to 90% of the buildout, making it dependent on credit markets.
AI Summary.At least 18 data-center financings in the junk-bond and leveraged loan markets carry double-digit yields, and a $5bn syndicated loan for neocloud Volta Infrastructure required steeper discounts and higher interest, pushing its yield above 11%. The average yield on investment-grade technology bonds has risen to about 6%, roughly 1.4 percentage points higher than a year earlier.
Neil Callanan, Caleb Mutua, and Paula DoeneckeBloomberg
Date Posted:
The bid-to-cover ratios for hyperscaler debt continue to decline while the cost of insuring against losses on the debt related to the AI buildout continues to rise. At least 18 data-center financings across the junk-bond and leveraged loan markets now carry double-digit yields.
Are investors demanding higher yields to finance data centers?
Core argument: At least 18 data-center financings in the junk-bond and leveraged loan markets carry double-digit yields, according to data compiled by Bloomberg.
At least 18 data-center financings across the junk-bond and leveraged loan markets now carry double-digit yields, according to data compiled by Bloomberg. Even new deals are feeling the strain. JPMorgan and Goldman Sachs had to offer steeper discounts and higher interest payments to drum up demand for neocloud Volta Infrastructure Holdings' $5 billion syndicated loan, pushing its yield above 11%. Nowhere is that unease more pronounced than in the CDS market. The cost of insuring Oracle debt, widely seen as a barometer of risk across the AI boom, reached a fresh record last week. So did default protection costs for SpaceX, Nvidia and Meta. The pressure is showing up in their borrowing costs. The average yield on bonds in the Bloomberg Investment Grade Technology Index has surged to about 6%, about 1.4 percentage points higher than a year ago.
Takeaways by Macro Roundup® AI
At least 18 data-center financings in the junk-bond and leveraged loan markets carry double-digit yields, according to data compiled by Bloomberg.
JPMorgan Chase and Goldman Sachs offered steeper discounts and higher interest payments to generate demand for neocloud Volta Infrastructure Holdings’ $5 billion syndicated loan, pushing its yield above 11%.
The cost of insuring Oracle’s debt, widely seen as a barometer of risk across the AI boom, reached a record last week, as did default protection costs for SpaceX, Nvidia and Meta.
Related Articles:
AI Debt Worries Spark Surge In Default Insurance Trading— Weekly trading in credit default swaps—insurance against bond default—tied to Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and Nvidia debt averaged over $9bn in a recent month, up from under $2bn, according to Morgan Stanley Research. Five-year protection on Meta and Nvidia reached record highs, while swaps on CoreWeave traded near distressed levels.
Hyperscaler CDS Widening Is Not a Dealer Inventory Story— The cost of insuring large artificial intelligence firms' debt against default has widened while bank spreads have stayed flat near 40 basis points, which the author argues rules out dealer hedging of new bond supply as the cause. Instead, markets are repricing a debt-financed AI investment cycle with rising leverage and negative free cash flow.
AI Set to be Largest CapEx Cycle Ever … and Soon Majority Externally Financed— Hyperscaler and data-center investment is tracking toward $1.5tn annually by 2030, which in real terms would exceed the 2005-06 housing boom, the 1990s fiber peak, and the 1882 railroad peak combined. External debt and equity are projected to fund close to 90% of the buildout, making it dependent on credit markets.
AI Summary.The cyclically adjusted price-earnings ratio for the S&P 500 stands at 40.7, implying an earnings yield of 2.46% and, combined with an estimated 1.8% annual real per-share earnings growth, a real expected return near 3% a year. Against a 2.75% inflation-protected 10-year Treasury yield, that leaves an equity premium of roughly 25 basis points, assuming the elevated multiple holds.
Hanno LustigThe Two Cents
Date Posted:
Is Database:
Database
At 40.7, CAPE is nearing its December 1999 peak of 44.2. For the S&P 500 to beat TIPS over the next decade, today’s near-record P/E multiple must persist for ten years, or real earnings must grow more than 5.8% a year—a pace exceeded in only two decades since 1880: the 1920s and 2010s.
How much do stock valuations offer over inflation-protected Treasuries?
Core argument: The S&P 500’s cyclically adjusted price-earnings ratio stands at 40.7, close to an all-time high, implying an earnings yield of just 2.46% for buyers of the index.
The CAPE is currently trading at 40.7, which is close to an all-time high. You’re paying close to 41 times annual earnings when you buy the S&P 500. The implied earnings yield is only 2.46%. According to AQR, a reasonable estimate for real earnings growth (on a per share basis) is 1.8% per annum. If you plug in today’s numbers, you end up with a real expected return of around 3% per year for U.S. equity. That’s a surprisingly low number. When you compare it to the 10-year TIPS yield of 2.75%, you’re left with a razor-thin equity premium. The stock market will only yield an extra 25 basis points per annum over the next, say, 10 years to compensate you for taking on equity risk. You might think this is a very conservative estimate, but it’s actually not, because we’re implicitly assuming that the CAPE stays at 40.7 over the next 10 years. That’s quite a lofty multiple. Even casual eyeball econometrics suggests that that’s not a great assumption, because the price-earnings ratio tends to revert back to its mean.
Takeaways by Macro Roundup® AI
The S&P 500’s cyclically adjusted price-earnings ratio stands at 40.7, close to an all-time high, implying an earnings yield of just 2.46% for buyers of the index.
Combining the 2.46% earnings yield with AQR’s estimate of 1.8% annual real per-share earnings growth yields a real expected return of roughly 3% per year for U.S. equity.
Against a 10-year TIPS yield of 2.75%, U.S. equities offer an equity premium of about 25 basis points per year, an estimate that assumes the CAPE stays at 40.7 for a decade despite the price-earnings ratio’s tendency to revert to its mean.
Related Articles:
Financing the AI Buildout— The authors estimate roughly 57 gigawatts of operating US data center capacity against a planned development pipeline of about 509 gigawatts, which is not a completion forecast since projects may be delayed, scaled back, or abandoned. Their central scenario has 182.8 gigawatts of added capacity becoming operational through 2032, 117.2 gigawatts completed later, and 226.9 gigawatts never completed.
A Catalyst For The AI Bubble Break— Historically, controlling for valuations, a 1% rise in the share of the market represented by initial public offerings has been associated with roughly a 4% decline in prices. With large artificial-intelligence company listings expected and share issuance exceeding buybacks, equity supply could shift from shrinking about 1% a year to growing near 5%, implying returns roughly 20% below average.
One Hundred Years in the U.S. Stock Markets— Over a century, U.S. publicly traded stocks generated roughly $91tn in shareholder wealth enhancement, with returns strongly positively skewed and dominated by a relatively small number of high-performing firms. Because most possible outcomes fall below the mean in a skewed distribution, the authors argue financial planning should focus on median outcomes or the full distribution.
AI Summary.The decline in US Treasury yields since their recent peak was driven by falling real yields rather than lower inflation expectations, which have been stable, a pattern consistent with a risk-off bond rally. Stock declines have been concentrated in hyperscalers, semiconductors, and data-centre-linked industrials, while nearly three-quarters of S&P 500 stocks are up.
Robert ArmstrongFinancial Times
Date Posted:
Armstrong argues that the Treasury rally, with the 10-year yield falling to 4.96% and an accompanying rout in technology stocks, has “the feeling of a turning point.” Since October 5, data-center-related sectors have sold off, while defensive sectors—health care and consumer staples—have outperformed.
Are falling Treasury yields signaling risk aversion rather than lower inflation?
Core argument: Treasury yields peaked in the first week of this month, and the entire decline since then came from falling real yields, with break-even inflation expectations extremely stable rather than reflecting expected disinflation.
Treasury yields peaked in the first week of this month. The decline since then does not reflect an anticipated reduction in inflation. The whole move was driven by falling real yields. Break-even inflation expectations have been extremely stable. A drop in real yields that is not about central bank positioning is usually about slowing growth expectations, or the perception of rising risks elsewhere in the markets. This looks like the latter — a classic risk-off bond rally, driven by fear. The pieces for this fear trade have been falling in pace for some time now. The stocks that have fallen most since yields peaked, on the other hand, are remarkably concentrated among hyperscalers, semiconductors and industrial stocks associated with data centre construction. Everyone else is doing OK; almost three-quarters of the stocks in the S&P 500 are up since October 5.
Takeaways by Macro Roundup® AI
Treasury yields peaked in the first week of this month, and the entire decline since then came from falling real yields, with break-even inflation expectations extremely stable rather than reflecting expected disinflation.
A fall in real yields unrelated to central bank positioning typically reflects slowing growth expectations or perceived rising risks elsewhere, and this move looks like a risk-off bond rally driven by fear.
Stock declines since yields peaked are concentrated among hyperscalers, semiconductors and industrial stocks tied to data centre construction, while almost three-quarters of S&P 500 stocks are up since October 5.
Related Articles:
Califoreboding— A basket of 42 AI stocks accounts for 60%-80% of all S&P 500 earnings growth, price returns and capital spending, while 60% of S&P 500 stocks sit 20% or more below their highs and only 30% outperform the index. AI-related construction spending has risen 38%, the only positive trend as manufacturing, commercial real estate and residential construction decline.
Meta Credit Risk Gauge Surges to Record on AI Spending Plans— The cost of insuring Meta Platforms' debt against default reached a record high and its bonds weakened as plans to increase capital spending revived investor concern that large technology firms will issue more debt to fund artificial intelligence buildouts. Default protection costs also rose for Oracle, Microsoft, and Broadcom.
Slowing Token Spend Shows Compute Demand Isn’t Unlimited— Demand for computing power appears highly price-sensitive, with modest increases in token prices pushing buyers to cut usage or switch to cheap open-source alternatives, undercutting claims that demand is unlimited. Return on invested capital at the five largest cloud providers is declining as invested capital grows faster than the earnings it generates.
AI Summary.An analysis indexes 10-year government bond yields across countries from the start of the year, cumulating their rise in basis points, and compares each against an equal-weighted average of daily changes in all those yields to show which countries are performing better or worse than average.
Robin BrooksRobin Brooks Substack
Date Posted:
Brooks argues a global debt crisis is underway, pointing to contagion from France driving rising yields in Italy and Spain, and a flight to safety driving down yields in the US, while low-debt safe havens like Germany, Sweden and Switzerland extend their outperformance.
Are government bond yields rising faster in some countries than others?
Core argument: The charts index 10-year government bond yields to 100 on January 1, 2026, and cumulate each country’s rise in basis points, so a reading of 200 means a yield up 100 basis points since the start of 2026.
The blue lines in the charts show how 10-year government bond yields have evolved so far this year. I’ve indexed them to 100 on January 1, 2026, and cumulate the rise in basis points from that day to yesterday. If you’re at 200 as of yesterday, that means your ten year yield is up 100 basis points since the start of 2026. The black line in each chart is the same and is the equal-weighted average of daily changes in all of these yields. I cumulate this up to show who’s better or who’s worse than average.
Takeaways by Macro Roundup® AI
The charts index 10-year government bond yields to 100 on January 1, 2026, and cumulate each country’s rise in basis points, so a reading of 200 means a yield up 100 basis points since the start of 2026.
A black line identical across the charts cumulates the equal-weighted average of daily changes in all the yields shown, serving as the benchmark for which countries are better or worse than average.
Related Articles:
What’s the Matter With France?— France's official retirement age for full benefits was 62 in 2023, versus 67 in Denmark and the US, and its average actual retirement age was the lowest in Western Europe while life expectancy at 65 reaches about 87, two years longer than in the US.
Global Debt Report 2026— Gross central government borrowing across OECD countries rose 7% to a record $17tn, about 23% of GDP, and is projected to reach roughly $18tn, or 24% of GDP. The United States and Japan account for close to 80% of the area's refinancing requirements, with the US share rising from 57% to 70%.
Are Government Bonds Safe in Times of War and Pandemic?— Government bonds severely underperform in wartime, in absolute terms and relative to equities, housing, and economic growth, unlike in financial crises when bonds tend to outperform stocks. The U.S. and U.K. reduced war-driven debt-to-GDP ratios without explicit default, using surprise inflation and financial repression that shifted fiscal burdens onto bondholders.
AI Summary.Payment-data analysis finds 80% of the largest corporate AI spenders are venture-backed, versus 13% of the median spender, indicating that much token spending is funded by venture capital rather than earnings. If AI payoffs are slow, venture returns could suffer and a funding pullback could weaken demand for computing power.
Torsten SlokApollo
Date Posted:
Is Database:
Database
According to Ramp data, 80% of firms in the top 1% of AI spending are VC-backed, compared with just 13% of firms at the median. This suggests that the largest purchasers of AI computing capacity rely heavily on venture capital rather than internally generated earnings.
Is venture capital funding most corporate AI token spending?
Core argument: Ramp data show 80% of the top 1% of AI spenders are venture-backed, versus 13% of the median spender, suggesting the largest buyers of compute are funding token spending with venture capital rather than organic earnings.
Ramp data show that 80% of the top 1% of AI spenders are VC-backed, compared with only 13% of the median spender, suggesting that the circularity seen among AI hyperscalers and model providers also extends to token spending, as the startups buying the most compute are paying for it with venture capital rather than organic earnings. If the AI payoff is slow to materialize or never arrives, VC returns would come under pressure, since the capital funding today’s token spending would have little to show for it, and a pullback in venture funding could in turn weaken demand for compute.
Takeaways by Macro Roundup® AI
Ramp data show 80% of the top 1% of AI spenders are venture-backed, versus 13% of the median spender, suggesting the largest buyers of compute are funding token spending with venture capital rather than organic earnings.
Ramp’s data on the venture backing of the heaviest AI spenders suggest the circularity seen among AI hyperscalers and model providers also extends to token spending.
If the AI payoff is slow to materialize or never arrives, venture capital returns would come under pressure, and a pullback in venture funding could in turn weaken demand for compute.
Related Articles:
Slowing Token Spend Shows Compute Demand Isn’t Unlimited— Demand for computing power appears highly price-sensitive, with modest increases in token prices pushing buyers to cut usage or switch to cheap open-source alternatives, undercutting claims that demand is unlimited. Return on invested capital at the five largest cloud providers is declining as invested capital grows faster than the earnings it generates.
AI Capex, Required Revenues, And What’s Priced In The AI Trade— Goldman Sachs projects the largest US hyperscalers' capital spending reaches $1.2tn in 2027 and $1.4tn in 2028, with growth decelerating, and estimates they need roughly $300bn in annual AI revenues to break even. AI infrastructure stocks have de-rated, with hyperscalers at their smallest recorded premium to the median S&P 500 company.
Financing the AI Buildout— The authors estimate roughly 57 gigawatts of operating US data center capacity against a planned development pipeline of about 509 gigawatts, which is not a completion forecast since projects may be delayed, scaled back, or abandoned. Their central scenario has 182.8 gigawatts of added capacity becoming operational through 2032, 117.2 gigawatts completed later, and 226.9 gigawatts never completed.
AI Summary.Weekly trading in credit default swaps—insurance against bond default—tied to Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and Nvidia debt averaged over $9bn in a recent month, up from under $2bn, according to Morgan Stanley Research. Five-year protection on Meta and Nvidia reached record highs, while swaps on CoreWeave traded near distressed levels.
Ramsay Hodgson and Euan HealyFinancial Times
Date Posted:
Since May, Nvidia’s 5-year CDS spread has more than doubled (~43 to 101bp), while Meta’s climbed from ~73 to 108bp. Both are at record highs. The North American IG CDX index rose only ~7bp (55 to 62bp), signalling a widening AI credit-risk premium.
Are investors buying more default protection on tech company debt?
Core argument: Weekly trading in credit default swaps linked to bonds from Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and Nvidia averaged more than $9bn in September, up from less than $2bn in July, according to Morgan Stanley Research.
Five-year CDS spreads on Meta and Nvidia have climbed to 108bps and 101bps, respectively, this month, both record highs. By far the greatest concern centres on neocloud CoreWeave, some of whose bonds have fallen into distressed territory. Five-year CDS on the company's debt is trading at about 1,090bps. A purchaser of $10mn of protection would need to pay $1.82mn upfront, plus quarterly payments of $125,000. More than $9bn worth of CDS linked to bonds issued by Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and Nvidia was traded on average each week in September, up from less than $2bn in July, according to Morgan Stanley Research. Trading volumes in US corporate CDS indices have already surpassed last year's total, according to Bloomberg data.
Takeaways by Macro Roundup® AI
Weekly trading in credit default swaps linked to bonds from Oracle, SpaceX, Alphabet, Amazon, Meta, Broadcom and Nvidia averaged more than $9bn in September, up from less than $2bn in July, according to Morgan Stanley Research.
Five-year CDS on Meta climbed to 108bp and on Nvidia to 101bp, both record highs, meaning insuring $10mn of Meta debt against default costs $108,000 a year.
Five-year CDS tied to neocloud CoreWeave trades at about 1,090bp, requiring $1.82mn upfront plus quarterly payments of $125,000 for $10mn of protection, with some of its bonds sold down into distressed territory.
Related Articles:
CoreWeave Risk Gauge Hits Record, Pricing 60% Odds of Default— The cost of insuring CoreWeave's debt against default over five years reached a record 1194.5 basis points, with these swaps pricing in a more than 60% chance of default within five years. Its 2031 bonds are among the biggest decliners in the US junk market, with yields pushed to 15.6% and 16.2%.
Hyperscaler CDS Widening Is Not a Dealer Inventory Story— The cost of insuring large artificial intelligence firms' debt against default has widened while bank spreads have stayed flat near 40 basis points, which the author argues rules out dealer hedging of new bond supply as the cause. Instead, markets are repricing a debt-financed AI investment cycle with rising leverage and negative free cash flow.
AI Set to be Largest CapEx Cycle Ever … and Soon Majority Externally Financed— Hyperscaler and data-center investment is tracking toward $1.5tn annually by 2030, which in real terms would exceed the 2005-06 housing boom, the 1990s fiber peak, and the 1882 railroad peak combined. External debt and equity are projected to fund close to 90% of the buildout, making it dependent on credit markets.