Edward Conard

Top Ten New York Times Bestselling Author

  • “Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
  • “…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 fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “…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 comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Tuesday, November 11, 2025

AI Summary. Investors have sold off bonds issued by large US technology companies building data centres, including Alphabet, Meta, Microsoft and Oracle, as concerns over artificial intelligence spending reached the bond market. Bank of America data show the extra yield investors demand to hold this debt over Treasuries rose to 0.78 percentage points from 0.5.

Kate Duguid and Tabby Kinder Financial Times
Date Posted:

“Jitters over Silicon Valley’s boom in spending on AI have spilled into the bond market.” According to BAC data, a basket of bonds issued by the major hyperscalers has seen an uptick in its spread over Treasuries to 0.78pp, up from 0.5pp in September.

Are bond investors demanding higher yields on tech AI spending debt?

Core argument: The spread on a basket of hyperscaler bonds over Treasuries has climbed to 0.78 percentage points from 0.5 points in September, the highest since April, when Donald Trump's tariff plans sent markets reeling, according to Bank of America data.

Investors have been selling off the debt of US tech heavyweights, showing how jitters over Silicon Valley’s boom in spending on artificial intelligence have spilled into the bond market. A basket of bonds issued by so-called hyperscalers — companies that are building vast data centres, including Alphabet, Meta, Microsoft and Oracle — has sustained a hit in recent weeks. The spread, or premium in yield investors demand to buy the debt over Treasuries, has climbed to 0.78 percentage points, the highest level since Donald Trump sent markets reeling in April with his tariff plans, and up from 0.5 points in September, according to Bank of America data.

Takeaways by Macro Roundup® AI

  1. The spread on a basket of hyperscaler bonds over Treasuries has climbed to 0.78 percentage points from 0.5 points in September, the highest since April, when Donald Trump's tariff plans sent markets reeling, according to Bank of America data.
  2. Investors have been selling the debt of US tech heavyweights including Alphabet, Meta, Microsoft and Oracle, carrying jitters over AI spending into the bond market.

Related Articles:

  • Meta, xAI Starting Trend for Billions in Off-Balance Sheet Debt — Off-balance-sheet structures such as special purpose vehicles and asset-tied joint ventures are becoming the standard financing route for AI data center deals. Morgan Stanley estimates tech firms will need as much as $800bn from private credit by 2028, with roughly $1.5tn of external financing required across the AI ecosystem.
  • Minsky Moments and AI CapEx — Capital-spending forecasts for the four largest cloud providers rose nearly 50% over the year, and AI data centers are increasingly funded through leveraged and off-balance-sheet structures, including $3.2bn in senior secured notes and a roughly $30bn asset-backed financing. The author argues this marks a speculative-finance phase where borrowers must roll over principal.
  • Four Things: OpenAI’s Backstop, CoreWeave CDS, Oregon, and Humanity’s TAM — A suggestion that the U.S. government could guarantee data-center financing, later disavowed by OpenAI's leadership, signaled that AI build-out costs have reached utility proportions. The author argues lenders may now price OpenAI more like a sovereign-backed utility than a growth company, exposing the fragility of the capital structure behind compute-intensive AI.

AI Summary. Japan's public sector operates as a highly leveraged sovereign wealth fund, funding risky assets by borrowing from banks and depositors at below-market rates, with returns sustaining its debt-to-GDP ratio above 200% without primary surpluses. The authors argue the resulting duration and currency mismatches expose Japan to mark-to-market losses that could trigger a fiscal crisis or a run on the yen.

Yili Chien, Wenxin Du, and Hanno Lustig Journal of Economic Perspectives
Date Posted:
Is Database:
Database

Japan has so far avoided inflation and debt crises despite decades of slow growth and deficits that left gov debt at > twice GDP by “running a risky sovereign wealth fund” with maturity and currency mismatch, financed at below-market interest rates.

How does Japan sustain debt above 200% of GDP without surpluses?

Core argument: Japanese government debt exceeds 200% of GDP with decades of budget deficits and sluggish growth, yet inflation has remained subdued and no major debt crisis has emerged.

Japan presents a striking puzzle in public finance. Government debt exceeds 200 percent of GDP, budget deficits have persisted for decades, and economic growth has been sluggish. Yet inflation has remained subdued, and no major debt crisis has emerged. To keep its promises to bondholders and to Japanese transfer recipients, the Japanese government has combined low-interest-rate policies with risk-taking on the asset side. The Japanese public sector is running a highly leveraged and increasingly risky sovereign wealth fund, exposing its taxpayers, bondholders, and depositors to significant interest rate, exchange rate, and equity risks. The risky investments are funded by borrowing from banks and ultimately depositors at below-market rates. This strategy has worked particularly well over the past two decades, amid a generally low-yield global macroeconomic environment. Returns on risky investments have sustained the debt-to-GDP ratio in Japan despite the lack of primary surpluses. Going forward, the sizable duration and currency mismatch on the public sector’s balance sheet exposes Japan to significant mark-to-market losses that could trigger a fiscal crisis and even a run on the Japanese currency. Unlike Japan, the US government has relied more on external funding to support its deficits.

Takeaways by Macro Roundup® AI

  1. Japanese government debt exceeds 200% of GDP with decades of budget deficits and sluggish growth, yet inflation has remained subdued and no major debt crisis has emerged.
  2. Japan's public sector operates a highly leveraged sovereign wealth fund, funding risky investments by borrowing from banks and ultimately depositors at below-market rates, exposing taxpayers, bondholders, and depositors to interest rate, exchange rate, and equity risks.
  3. Returns on Japan's risky public-sector investments have sustained its debt-to-GDP ratio despite the absence of primary surpluses, while the duration and currency mismatch exposes Japan to mark-to-market losses that could trigger a fiscal crisis and even a run on the currency.

Related Articles:

  • Japan’s Debt Crisis — The Bank of Japan's holdings of Japanese government bonds are so large that maturing bonds exceed gross purchases, making it a net seller equal to -2.6% of GDP over a recent one-year period. Further cuts to gross buying would deepen the decline in its holdings and likely push longer-term Japanese yields higher.
  • Global Debt Report 2025 — About 60% of fixed-rate government debt in advanced economies maturing by 2027 — roughly $9tn — was issued before the recent interest-rate tightening, with weighted average yields below 2% versus projected 10-year rates near 3.6%, implying refinancing at nearly double the original cost. France, Spain, the UK, and the US face heightened vulnerability, with maturing debt exceeding 15% of GDP.
  • Should We Be Alarmed Or Optimistic About Japan’s Debt? — Japan's 30-year government bond yield reached an all-time high of 3.21% following weak auctions, while Morgan Stanley reports the budget deficit was nearly eliminated in the first quarter, the best public finances in almost three decades. The return of nominal wage and price growth lifted income and consumption tax revenues, offset by lower household savings.

AI Summary. Estimates comparing the 10-year US Treasury yield with currency-hedged 10-year government bond yields in other advanced economies show the US "convenience yield" eroding steadily over the past decade, with the premium the US pays on 10-year Treasuries shrinking more recently. Sweden's convenience yield has fallen below Switzerland's, making it the top safe-haven destination on this measure.

Robin Brooks Robin Brooks Substack
Date Posted:

The 10-year Treasury yield, relative to the trade-weighted, exchange rate-adjusted yield on 10-year governments in other advanced economies, rose over the last decade but has fallen in recent months, suggesting “many other fiscal train wrecks out there.”

Is the US losing its safe-haven advantage in government bonds?

Core argument: Brooks estimates the US "convenience yield" — the 10-year Treasury yield versus trade-weighted 10-year yields in other advanced economies swapped into US dollars via FX forwards — has eroded steadily over the past decade, a move predating the second Trump administration.

[Brooks updated his] “convenience yield” estimates. For the US, this calculation compares the 10-year Treasury yield to the trade-weighted yield on 10-year government bonds in other advanced economies after those are swapped back into US Dollars using FX forwards. [There have been] two key developments. First, the US “convenience yield” has eroded steadily over the past decade, a move that predates the second Trump administration [left chart]. But - surprisingly - it looks like the premium the US pays on 10-year Treasuries has shrunk in recent months, which [Brooks] interpret as a sign that there’s many other fiscal trainwrecks out there, so markets think the US looks ok on a relative basis [right chart]. Second, Sweden is emerging as the top safe haven destination, with a “convenience yield” that just fell below Switzerland.

Takeaways by Macro Roundup® AI

  1. Brooks estimates the US "convenience yield" — the 10-year Treasury yield versus trade-weighted 10-year yields in other advanced economies swapped into US dollars via FX forwards — has eroded steadily over the past decade, a move predating the second Trump administration.
  2. The premium the US pays on 10-year Treasuries has shrunk in recent months, which Brooks interprets as a sign that many other fiscal trainwrecks exist and markets view the US as acceptable on a relative basis.
  3. Sweden is emerging as the top safe haven destination in Brooks's estimates, with a "convenience yield" that just fell below Switzerland's.

Related Articles:

  • Across The Rich World, Fiscal Crises Loom — An analysis estimates most rich countries could keep debt stable as a share of their economies with only small primary surpluses—Britain's required surplus is 0.3% of GDP—even if all debt were refinanced immediately at current rates, but the tightening needed to stabilize debt ratios exceeds 2% of GDP in Britain and America and 3% in France.
  • The Global Debt Overhang — Across much of the G10, 30-year government bond yields have risen even as central banks cut rates, which the author reads as a sign markets are on average becoming more averse to holding government debt. Switzerland is the only country whose 30-year yield fell, while Germany and the UK stand out as disproportionate risers.
  • Global Debt Report 2025 — About 60% of fixed-rate government debt in advanced economies maturing by 2027 — roughly $9tn — was issued before the recent interest-rate tightening, with weighted average yields below 2% versus projected 10-year rates near 3.6%, implying refinancing at nearly double the original cost. France, Spain, the UK, and the US face heightened vulnerability, with maturing debt exceeding 15% of GDP.

AI Summary. Bank of America deposit and spending data indicate nearly 24% of households in 2025 spend more than 95% of income on necessities such as housing, groceries, utilities and childcare, a 0.3 percentage point increase from the prior year, with growth nearly 3x slower.

Joe Wadford Bank of America
Date Posted:
Is Database:
Database

BAC internal data show the share of households devoting more than 95% of household income to “necessity spending” increasing 0.3pp to 24% in the first 3 quarters of 2025 compared to the same period in 2024 – ~⅓ of the previous year’s pp rise.

Is the share of households living paycheck to paycheck still rising?

Core argument: Bank of America estimates nearly 24% of households in 2025 are living "paycheck to paycheck," defined as necessity spending exceeding 95% of household income, up 0.3 percentage points from 2024 at a growth rate nearly three times lower than the prior year.

Inflation has stayed fairly high, at 3.0% year-over-year in September, and after-tax wage and salary growth, based on Bank of America deposit data, for middle- and lower-income households has not kept pace, increasing 2% and 1% YoY, respectively, in October. One way to measure how slowing wages and sticky inflation levels are pressuring consumers’ wallets is to estimate the proportion of households living “paycheck to paycheck,” using Bank of America internal consumer deposit and spending data. We define living “paycheck to paycheck” using a fairly broad definition of necessity spending, covering areas such as housing, gasoline, groceries, utility bills, internet service provider subscriptions, public transportation, and childcare. We then identify the proportion of households in which necessity spending is more than 95% of their household income, leaving them relatively little or nothing left over for “nice-to-have” discretionary spending or savings. We use this as our central indicator. Exhibit 2 shows that nearly 24% of households so far in 2025 would be classified as living “paycheck to paycheck” based on our central indicator. This is a 0.3 percentage point increase on 2024, but a growth rate nearly three times lower than last year. The share of households where necessity spending is more than 90% of their household income moved similarly.

Takeaways by Macro Roundup® AI

  1. Bank of America estimates nearly 24% of households in 2025 are living "paycheck to paycheck," defined as necessity spending exceeding 95% of household income, up 0.3 percentage points from 2024 at a growth rate nearly three times lower than the prior year.
  2. After-tax wage and salary growth for middle- and lower-income households rose 2% and 1% year-over-year respectively in October, trailing inflation of 3.0% year-over-year in September, based on Bank of America deposit data.
  3. Bank of America's necessity spending definition covers housing, gasoline, groceries, utility bills, internet service provider subscriptions, public transportation, and childcare, and the share of households spending more than 90% of income on necessities moved similarly to the 95% measure.

Related Articles:

  • Real Income Growth Shifts Down, Especially For The Young — Inflation-adjusted median income growth for 25–54-year-old workers declined and has hovered near decade lows, with a more marked slowdown among those aged 25–29. The authors suggest this may reflect a reduced pace of job-to-job transitions, seen in lower hiring and quit rates.
  • Jobless Growth — Labor demand growth may be too weak to keep pace with even modest labor supply growth: the labor market is somewhat weaker than before the pandemic, job growth has turned negative outside healthcare, and firms are focused on using AI to cut labor costs. Economists estimate most employment losses in routine manual occupations since the 1970s occurred during recessions.
  • Don’t Choose Your Own Adventure: Understanding Middle-Class Earnings Trends — Using a standard price index, US men's earnings were flat from 1973 to 1989 and rose 0.4% a year thereafter, while an alternative middle-class cost-of-living index shows 0.7% annual growth in both periods and a 40% total rise instead of 13%. The author attributes weak growth to 1970s stagflation and early-1980s recessions rather than trade, immigration, or income concentration.

AI Summary. In New York City's mayoral race, Mamdani's average margin was 30 points in precincts with a median age under 45, and voters under 45 made up 43% of in-person ballots versus 36% in the prior governor's race. He won by over 50 points in neighborhoods where the white population share grew and the Black or Hispanic share declined.

Maya King, Jeff Adelson and Alex Lemonides New York Times
Date Posted:
Is Database:
Database

Voters under the age of 45, 46% of registered voters in New York City, made up ~43% of voters in the mayor’s race. In neighborhoods where the nonwhite population is shrinking, Mamdani won by more than 50pp.

How did age and neighborhood demographic shifts shape a mayoral victory?

Core argument: Mamdani's average margin of victory reached 30 points in precincts where the median age is under 45, among the highest of any demographic group.

In precincts where the median age is under 45, [Mamdani's] average margin of victory was 30 points — among the highest of any demographic group. Voters younger than 45 — about 46% of registered voters in the city — made up about 43% of those who cast their ballots in person early or on Election Day. That outpaces the 2022 midterms, when Gov. Kathy Hochul, a Democrat, narrowly fought off a challenge from Lee Zeldin. That year, younger voters made up 36% of those who voted. (Mail ballots, which are still being accounted for in the 2025 election, generally make up a small share of overall votes.) In neighborhoods where the nonwhite population is shrinking [Mamdani] notched some of his strongest performances. In parts of the city where the Black or Hispanic population share had declined and the white population share had grown since before the pandemic, Mr. Mamdani won by more than 50 percentage points.

Takeaways by Macro Roundup® AI

  1. Mamdani's average margin of victory reached 30 points in precincts where the median age is under 45, among the highest of any demographic group.
  2. Voters younger than 45, about 46% of New York City's registered voters, cast about 43% of in-person early and Election Day ballots, up from 36% in the 2022 midterms, when Kathy Hochul narrowly fought off Lee Zeldin.
  3. Mamdani won by more than 50 percentage points in parts of the city where the Black or Hispanic population share had declined and the white population share had grown since before the pandemic.

Related Articles:

  • As New Jobs In Finance Dry Up, New York City’s Fiscal Model Is Wilting — New York state's share of US taxpayers reporting over $1m in income fell from 12.7% in 2010 to 8.7% in 2022, and the Citizens Budget Commission estimates state and city income tax receipts from that group would have been $13bn higher had the share held. Job growth has concentrated in lower-paid health care and social assistance.
  • Zero-Sum Thinking and the Roots of U.S. Political Divides — Zero-sum thinking — viewing gains for one group as losses for another — is associated with greater support for liberal economic policies, stronger alignment with the Democratic Party, and weaker alignment with the Republican Party. The authors also find it linked to believing the QAnon conspiracy theory holds some truth and to sympathy for the attack on the U.S. Capitol.
  • Why is New York City Shrinking? — New York City's population is shrinking as domestic out-migration has accelerated, with those leaving tending to be younger, higher-income, and with fewer children than the city's overall population. Census Bureau surveys that ask movers their reasons indicate jobs matter more than any other consideration, including housing affordability.

AI Summary. Studies of Quebec's shift to universal child care found increases in aggression, anxiety and hyperactivity among children and declines in motor and social skills, with large estimated effects. Researchers attribute the contrast with successful small randomized trials to slipping quality when a pilot is scaled up while holding down costs.

Economist Staff The Economist
Date Posted:
Is Database:
Database
Is Important:
Important

Baker, et al found that a shift to universal child care led to a rise in aggression, anxiety and hyperactivity among Quebecer children. James Heckman, who found very high returns to a small, high quality pilot program, warns against low-cost imitations.

Does scaling up universal child care harm children's development?

Core argument: Two studies by Baker, Gruber, and Mulligan found that Quebec's shift to universal child care appeared to raise aggression, anxiety and hyperactivity among children and lower motor and social skills, with large effects.

Two studies by Baker, Gruber, and Mulligan found that shifting to universal child care appeared to lead to a rise in aggression, anxiety and hyperactivity among Quebecer children, as well as a fall in motor and social skills, [and] the effects were large. How can this be? There is, after all, lots of evidence—in the form of small, randomised controlled trials—in support of similar schemes. [e.g.] In 2010, James Heckman and co-authors calculated that, on conservative assumptions, the investment made in the children [in the famous Perry study] returned 7-10% a year to society—comparable to the historical rate of return on equity. Quebec had tried, while keeping costs down, to imitate a small pilot on a grand scale. Standards inevitably slipped. When Heckman was asked by the NYT about the province’s experience, he noted how different it was from the trials he had studied. “These were warehouses,” he said. “They were fairly impersonal, there wasn’t any real quality. Quality has to be a sine qua non of the whole enterprise.”

Takeaways by Macro Roundup® AI

  1. Two studies by Baker, Gruber, and Mulligan found that Quebec's shift to universal child care appeared to raise aggression, anxiety and hyperactivity among children and lower motor and social skills, with large effects.
  2. James Heckman and co-authors calculated in 2010 that, on conservative assumptions, the investment in children in the Perry study returned 7–10% a year to society, comparable to the historical rate of return on equity.
  3. Heckman told the New York Times that Quebec's child-care settings were "warehouses" that were fairly impersonal and lacked real quality, unlike the trials he had studied, and argued that quality has to be a sine qua non.

Related Articles:

  • The Long-Run Impacts of a Universal Child Care Program — A universal childcare program in Quebec raised childcare use by over 40% of the baseline rate but is associated with lasting declines in noncognitive outcomes such as anxiety and aggression, persisting into the school years. The authors estimate boys saw the largest deterioration, with criminal accusations up 19% and convictions up 22%, while cognitive test score effects were mixed.
  • Universal Childcare, Maternal Labor Supply and Family Well-Being — The authors find Quebec's universal childcare raised the share of children aged 0-4 in care by 14 percentage points relative to the rest of Canada and married women's labor force participation by 7.7 percentage points. Tax revenue from added maternal work fell short of subsidy costs, and child and parental outcomes worsened.
  • Birth Rates Are Falling In The Nordics. Are Family-Friendly Policies No Longer Enough? — Falling Nordic birth rates are now driven by childlessness rather than smaller families, with roughly three-quarters of Finland's recent fertility decline attributable to people who have no children. Nearly 40% of Finnish men with low education are childless at 45, and most have no partners.

Wednesday, November 12, 2025

AI Summary. Global electricity demand is projected to rise roughly 40–50% by 2035, even though electricity currently accounts for only 21% of total final energy consumption worldwide. Electricity use by servers optimized for artificial intelligence is projected to rise fivefold by 2030, yet data centres are projected to account for less than 10% of global electricity demand growth through 2030.

Stéphanie Bouckaert, Christophe McGlade, Brent Wanner et al. International Energy Agency
Date Posted:
Is Database:
Database

Global electricity demand is forecast to rise by 40–50% by 2035. Only 10% of this growth is due to AI. Coal demand, driven by power generation in India and China, is 2x its 2000 level, but will likely peak before 2035 as cleaner energy comes online.

How much will global electricity demand rise by 2035?

Core argument: The World Energy Outlook 2025 projects electricity demand rising around 40–50% by 2035, though electricity accounts for only 21% of total final consumption while serving sectors making up over 40% of the global economy.

Electricity rises by around 40% to 50% by 2035. End-use electrification already accounts for half of today’s global energy investment. For the moment, electricity accounts for only 21% of total final consumption globally, but it is the key source of energy for sectors accounting for over 40% of the global economy and the main source of energy for most households. Driven by surging data centre expansion, electricity consumption by AI-optimised servers increases fivefold by 2030, due to a doubling of total electricity consumption by data centres by 2030. [Yet] data centres account for less than 10% of global electricity demand growth between 2024 and 2030. Cooling is a rising source of electricity demand, led by developing economies. In [one scenario] income-driven air conditioning use adds around 330 GW to global peak demand by 2035, and higher temperatures add another 170 GW. Since 2000, coal has been the fastest growing fossil fuel in the global energy mix. Global coal demand has nearly doubled since then, and reached the record level of 6,090 million tonnes of coal equivalent (Mtce) in 2024, driven mainly by the use of coal to generate power in emerging market and developing economies, in particular China and India. Global coal demand peaks before 2030 and by 2035, demand is around 8% lower than 2024 levels.

Takeaways by Macro Roundup® AI

  1. The World Energy Outlook 2025 projects electricity demand rising around 40–50% by 2035, though electricity accounts for only 21% of total final consumption while serving sectors making up over 40% of the global economy.
  2. Electricity consumption by AI-optimised servers increases fivefold by 2030 as total data centre electricity use doubles, yet data centres account for less than 10% of global electricity demand growth between 2024 and 2030.
  3. Global coal demand nearly doubled since 2000 to a record 6,090 million tonnes of coal equivalent in 2024, driven mainly by power generation in emerging market and developing economies, particularly China and India, and peaks before 2030, sitting about 8% below 2024 levels by 2035.

Related Articles:

  • Economic Growth Now Depends on Electricity, Not Oil — Average annual electricity load growth reached 3% and is projected to continue at that rate for five years, with utilities raising expected peak demand for 2029 by 12%, or 101 gigawatts, driven by data centers, new factories, electric vehicles, heat pumps, and hydrogen production.
  • Heliocentrism — An energy use must be electrified before green grid electricity can decarbonize it, and while US grid decarbonization continues at a steady pace, the electricity share of final US energy consumption has grown little. Transmission line growth runs well below Department of Energy targets for 2030 and 2035, and transformer delivery times have stretched from 4–6 weeks to 2–3 years.
  • What’s Happening to Wholesale Electricity Prices? — Average US retail electricity prices have risen about 35% since 2020, while average wholesale electricity costs have nearly doubled, with wholesale prices far more volatile. Wholesale location marginal prices consist of energy generation costs (typically 80–90% of the total), transmission congestion, and transmission losses, with a price spike partly driven by a large rise in natural gas prices.

AI Summary. Global data center, AI infrastructure and related power buildout could cost over $5tn, requiring funding from every public capital market plus private credit, alternative capital providers and possibly governments. The largest cloud providers generate over $700bn of annual operating cash flow and reinvest roughly $500bn in capital spending, with about $300bn a year estimated to fund AI and data centers.

Tarek Hamid, C. Stephen Tusa, Jr., Harlan Sur, et al. J.P. Morgan
Date Posted:
Is Database:
Database

JPM estimates that to meet an IRR hurdle of 10% on their forecast AI investment through 2030 would require ~$650B in annual revenue in perpetuity. That would be 58bps of global GDP, or $34.72/month from every current iPhone user.

How will a five-trillion-dollar AI infrastructure buildout be financed?

Core argument: Global data center, AI infrastructure and related power buildout could cost over $5 trillion in the authors' view, requiring funding from every public capital market plus private credit, alternative capital providers and government involvement.

Building out global data center and AI infrastructure and related power supplies could cost over $5 trillion in our view. Funding that extraordinary growth will likely require participation from every public capital market as well as private credit, alternative capital providers and even government involvement. Hyperscaler cash flow will have to do an enormous amount of the heavy lifting. Hyperscalers are generating over $700 billion of operating cash flow per annum, and reinvesting ~$500 billion of that back into capex. Note that ~$500 billion of reinvestment back into the collective businesses does not include ~$250 billion of research and development spending per year. We are assuming approximately $300 billion per year of hyperscaler capex reflects cash flow funded AI and data center capex (while AI is crowding out a lot of spending, it's not crowding out all spending). To drive a 10% return on our modeled AI investments through 2030 would require ~$650 billion of annual revenue into perpetuity, which is an astonishingly large number. But for context, that equates to 58bp of global GDP, or $34.72/month from every current iPhone user, or $180/month from every Netflix subscriber.

Takeaways by Macro Roundup® AI

  1. Global data center, AI infrastructure and related power buildout could cost over $5 trillion in the authors' view, requiring funding from every public capital market plus private credit, alternative capital providers and government involvement.
  2. Hyperscalers generate over $700 billion of operating cash flow per year and reinvest roughly $500 billion into capex, separate from about $250 billion of annual research and development spending, with roughly $300 billion of that capex assumed to fund AI and data centers.
  3. Earning a 10% return on the modeled AI investments through 2030 would require about $650 billion of annual revenue into perpetuity, equal to 58bp of global GDP, $34.72 per month from every current iPhone user, or $180 per month from every Netflix subscriber.

Related Articles:

  • Investor Angst Over Big Tech’s AI Spending Spills Into Bond Market — Investors have sold off bonds issued by large US technology companies building data centres, including Alphabet, Meta, Microsoft and Oracle, as concerns over artificial intelligence spending reached the bond market. Bank of America data show the extra yield investors demand to hold this debt over Treasuries rose to 0.78 percentage points from 0.5.
  • Meta, xAI Starting Trend for Billions in Off-Balance Sheet Debt — Off-balance-sheet structures such as special purpose vehicles and asset-tied joint ventures are becoming the standard financing route for AI data center deals. Morgan Stanley estimates tech firms will need as much as $800bn from private credit by 2028, with roughly $1.5tn of external financing required across the AI ecosystem.
  • Minsky Moments and AI CapEx — Capital-spending forecasts for the four largest cloud providers rose nearly 50% over the year, and AI data centers are increasingly funded through leveraged and off-balance-sheet structures, including $3.2bn in senior secured notes and a roughly $30bn asset-backed financing. The author argues this marks a speculative-finance phase where borrowers must roll over principal.

AI Summary. Technology valuations are not extreme by historical standards, with the Nasdaq Composite's average forward price-to-earnings ratio near 29 versus above 32 before the artificial intelligence boom. Concern instead centers on leverage: AI-related companies issued $139bn of corporate bonds this year, about 9% of investment-grade issuance and 23% above last year.

Asa Fitch and Dan Gallagher Wall Street Journal
Date Posted:

The share prices of AI-related firms as a multiple of forward earnings have seen a downtick as some investors worry about rising leverage. AI-related companies issued $139B of debt so far this year, or 9% of all investor-grade issuance.

Is rising debt, not valuations, the main risk in the AI boom?

Core argument: AI-related companies had issued $139 billion of corporate bonds this year as of last month, about 9% of all investment-grade issuance and a 23% jump over last year, according to a Goldman Sachs report.

Overall, tech valuations don’t look outrageous, especially with companies like Nvidia minting profits. The average forward price-to-earnings ratio for the tech-heavy Nasdaq Composite Index is about 29—a high number, to be sure, but nothing astronomical for a basket of growing companies. That figure reached above 32 in 2021, the year before the AI boom. [A] source of concern lies in the huge amount of leverage needed to finance the outsize ambitions of AI’s biggest players. Oracle reached a $300 billion deal with OpenAI in September to supply it with AI computing power, and it raised $18 billion through a bond sale the same month. That is a foretaste of what analysts say will be a growing pool of debt tied to data centers. AI-related companies had issued $139 billion of corporate bonds this year as of last month, according to a Goldman Sachs report—about 9% of all investor-grade issuance and a 23% jump over last year. Meta Platforms organized a $27.3 billion financing for a Louisiana data-center complex last month alone in what was the largest-ever private-debt deal.

Takeaways by Macro Roundup® AI

  1. AI-related companies had issued $139 billion of corporate bonds this year as of last month, about 9% of all investment-grade issuance and a 23% jump over last year, according to a Goldman Sachs report.
  2. The Nasdaq Composite's average forward price-to-earnings ratio is about 29, below the above-32 level reached in 2021, the year before the AI boom.
  3. Oracle reached a $300 billion deal with OpenAI in September to supply AI computing power and raised $18 billion through a bond sale that month, while Meta Platforms organized a $27.3 billion financing for a Louisiana data-center complex, the largest-ever private-debt deal.

Related Articles:

  • Why We Are Not In A Bubble….Yet — The 5 largest US technology companies are worth about 16% of the global public equity market, and the top 10 US stocks nearly 25%, almost $25tn. The author argues valuations are high but not bubble-level, with the "Magnificent 7" median forward price-to-earnings ratio at 27x, roughly half that of the late-1990s leaders.
  • Four Things: OpenAI’s Backstop, CoreWeave CDS, Oregon, and Humanity’s TAM — A suggestion that the U.S. government could guarantee data-center financing, later disavowed by OpenAI's leadership, signaled that AI build-out costs have reached utility proportions. The author argues lenders may now price OpenAI more like a sovereign-backed utility than a growth company, exposing the fragility of the capital structure behind compute-intensive AI.
  • Investor Angst Over Big Tech’s AI Spending Spills Into Bond Market — Investors have sold off bonds issued by large US technology companies building data centres, including Alphabet, Meta, Microsoft and Oracle, as concerns over artificial intelligence spending reached the bond market. Bank of America data show the extra yield investors demand to hold this debt over Treasuries rose to 0.78 percentage points from 0.5.

AI Summary. Average net debt relative to earnings for about 2,500 companies in the MSCI ACWI global stock index improved to 1.74x in the first half of this year from 2.53x a decade earlier, indicating stronger corporate balance sheets. Meanwhile, the International Monetary Fund projects government debt across major economies will rise annually through 2030 to about 137% of GDP.

Finbarr Flynn, Tasos Vossos, and Greg Ritchie Bloomberg
Date Posted:
Is Database:
Database

The debt of several blue-chip firms, including Microsoft in the US, Siemens in Germany and LVHM in France, trades at lower yields than the underlying government benchmark, as “tough fiscal compromises” look unlikely.

Are corporate balance sheets now stronger than government finances?

Core argument: Net debt to EBITDA for roughly 2,500 companies in the MSCI ACWI Index improved to an average of 1.74 times in the first half of this year from 2.53 a decade ago, according to data compiled by Bloomberg.

A key measure of corporate creditworthiness — net debt relative to earnings before interest, taxes depreciation and amortization — has hovered near its lowest post-financial-crisis level in recent years, despite higher borrowing costs. For about 2,500 companies in the MSCI ACWI Index, that metric improved to 1.74 times on average in the first half of this year from an average of 2.53 a decade ago, data compiled by Bloomberg show. A lower reading signals stronger balance sheets. By contrast, the debt-to-gross domestic product ratio across major economies has been going up, and the International Monetary Fund estimates it’ll rise every year through 2030 to about 137%. Microsoft has such strong earnings that its net debt is just a 10th of the last 12 months’ earnings. And in France, Orange — formerly known as France Telecom and a classic case of corporate deleveraging — is being rewarded by traders with a lower risk premium. Around 5% of French investment-grade notes trade at yields less than government debt, data compiled by Bloomberg show.

Takeaways by Macro Roundup® AI

  1. Net debt to EBITDA for roughly 2,500 companies in the MSCI ACWI Index improved to an average of 1.74 times in the first half of this year from 2.53 a decade ago, according to data compiled by Bloomberg.
  2. The International Monetary Fund estimates the debt-to-GDP ratio across major economies will rise every year through 2030 to about 137%, while corporate leverage has hovered near its lowest post-financial-crisis level.
  3. About 5% of French investment-grade notes trade at yields below government debt, according to data compiled by Bloomberg, and traders reward Orange, a case of corporate deleveraging, with a lower risk premium.

Related Articles:

  • Why Microsoft Has Lower Borrowing Costs Than the U.S. — Two AAA-rated Microsoft bonds trade at yields slightly below equivalent U.S. Treasurys, an unusual negative spread given that companies are more likely to default than a government able to tax and that corporate bonds are harder to sell. Demand for high-quality corporate debt has compressed the AAA corporate spread to 0.3 percentage point over Treasurys.
  • French Companies’ Borrowing Costs Fall Below Government’s As Debt Fears Intensify — Bond yields for French companies including L'Oréal, Airbus and Axa have fallen below those on French government debt of similar maturity, with 10 French firms trading at a negative spread to the sovereign — the most in data going back to 2006, according to Goldman Sachs. Sovereign yields typically set a baseline for corporate bonds.
  • Our Thoughts on Large US Deficits and Their Impact on Bond Yields — Total borrowing in the US is about average despite high government deficits, as high public borrowing has contributed to the need for higher interest rates that have depressed private sector borrowing. Risk premiums on long-dated debt look about average, with no market signs that economic or debt limits are currently being hit.

AI Summary. German industrial production stands at its 2005 level, with analysts arguing that core strengths—a hard-to-decarbonise industrial base, export dependence, and combustion-engine expertise—have become vulnerabilities. Chinese capital goods average 30% cheaper than European equivalents and have closed the quality gap, leaving Germany with its first capital-goods trade deficit with China since records began in 2008.

Olaf Storbeck, Sebastien Ash and Florian Müller Financial Times
Date Posted:
Is Database:
Database

German industrial production is only at its 2005 level, even after a partial rebound in September. Since the start of the year, Germany has run a deficit in capital goods with China, the first since 2008.

Have German industry's core strengths turned into vulnerabilities?

Core argument: German industrial production sits at its 2005 level even after a partial rebound in September, with Marcus Berret of Roland Berger arguing that many of Germany's economic core strengths have turned into vulnerabilities.

Industrial production sits at the 2005 level even after a partial rebound in September. “Many of Germany’s economic core strengths have turned into vulnerabilities,” says Marcus Berret, global managing director at Roland Berger, a Munich-based consultant. Those include a large industrial base that is hard to decarbonise, a high dependence on exports at a time when globalisation is under threat, and a mighty auto industry having to write off 140 years of internal combustion engine expertise. All this is exacerbated by two distinct political decisions in the US and China, taken 10 years apart: Donald Trump’s trade war, and Beijing’s decision a decade ago to turn itself into a global high-tech engineering powerhouse. On average, Chinese capital goods are 30% cheaper than those of Europeans. Crucially, manufacturers in the Asian superpower have also closed the quality gap. Since the start of 2025, Germany is now running a trade deficit in capital goods with China over a rolling 12-month period. That is a first since records began in 2008. Chinese machinery exports to Europe roughly doubled to around €40bn in over six years and may reach €50bn this year, according to industry association VDMA.

Takeaways by Macro Roundup® AI

  1. German industrial production sits at its 2005 level even after a partial rebound in September, with Marcus Berret of Roland Berger arguing that many of Germany's economic core strengths have turned into vulnerabilities.
  2. Chinese capital goods are on average 30% cheaper than European equivalents and Chinese manufacturers have closed the quality gap, leaving Germany with a rolling 12-month trade deficit in capital goods with China since the start of 2025, a first since records began in 2008.
  3. Chinese machinery exports to Europe roughly doubled to about €40bn over six years and may reach €50bn this year, according to industry association VDMA.

Related Articles:

  • Europe is Pricing Itself Out of Existence — Europe paid 166.3% more for its primary energy mix than it would have using conventional sources, equal to 5.6% of GDP rather than 3.37%, which the author argues largely explains minimal European growth. European Union primary energy consumption has fallen 17.47% from its peak, with renewables at 16.6% of the mix and renewable electricity priced at 5x conventional.
  • The Future of European Competitiveness – A Competitiveness Strategy for Europe — A European Union competitiveness report attributes the EU's labour productivity growth gap with the US largely to weaker activity in information and communications technology and large-scale digital services; excluding main ICT sectors, EU productivity was broadly at par with the US from 2000-2019. It estimates Europe's investment share must rise about 5 percentage points of GDP.
  • Europe Is Losing — Europe's share of global stock market valuations has fallen from 18% to 10% for the EU and from 8.3% to 2.6% for the UK, while America's has held at 48%. Tax revenue equals 38–44% of output in Germany, Italy and France versus 25% in the US, and industrial electricity costs 3–4x US levels.

AI Summary. Shortages of advanced semiconductors in China are acute enough that authorities have begun intervening in how output from Semiconductor Manufacturing International, the country's largest contract chip maker, is distributed, prioritizing Huawei's artificial-intelligence chip production. Chinese firms also smuggle Nvidia chips or access them remotely through cloud computing abroad.

Lingling Wei, Amrith Ramkumar and Robbie Whelan Wall Street Journal
Date Posted:

The Chinese government has started managing SMIC’s chip output, prioritizing Huawei’s demand as American export controls have created an “acute” shortage of advanced semiconductors in China.

Are chip export controls creating real semiconductor shortages in China?

Core argument: Shortages of advanced semiconductors in China are acute enough that the government has begun intervening in how output from Semiconductor Manufacturing International, the country's largest contract chip maker, is distributed, according to people familiar with the matter.

Shortages of advanced semiconductors are so acute that the government has begun intervening in how the output of China’s largest contract chip maker, Semiconductor Manufacturing International, is distributed, according to people familiar with the matter. Chinese authorities are trying to give priority to the needs of tech conglomerate and national champion Huawei Technologies, which uses SMIC technology to make artificial-intelligence chips, the people said. Companies in China also continue to smuggle Nvidia chips or access them remotely in other countries using cloud computing. According to contracts reviewed by The Wall Street Journal, at least 16 refrigerator-sized Blackwell racks—shipped in smaller parts and reassembled in China—were set to be delivered by November. That amount of chips isn’t sufficient to train advanced, large-scale AI models, but it can still be useful for conducting research and developing powerful applications.

Takeaways by Macro Roundup® AI

  1. Shortages of advanced semiconductors in China are acute enough that the government has begun intervening in how output from Semiconductor Manufacturing International, the country's largest contract chip maker, is distributed, according to people familiar with the matter.
  2. Chinese authorities are trying to give priority to Huawei Technologies, which uses SMIC technology to make artificial-intelligence chips, according to people familiar with the matter.
  3. At least 16 refrigerator-sized Blackwell racks, shipped in smaller parts and reassembled in China, were set to be delivered by November, according to contracts reviewed by The Wall Street Journal, an amount insufficient to train advanced large-scale AI models but useful for research and developing powerful applications.

Related Articles:

  • China Seeks To Triple Output Of AI Chips In Race With The US — China is preparing three new fabrication plants dedicated to producing Huawei's artificial intelligence processors, and people familiar with the effort say their combined capacity, once fully ramped up, could exceed current output from comparable lines at SMIC, China's leading chipmaker. Huawei denies having plans to launch its own fabrication plants.
  • China Has Its Own Concerns About Nvidia H20 Chips — China's leaders cite security concerns, including suspicion of hidden back doors in US-designed chips, as a rationale for their drive toward chip self-sufficiency. The dilemma extends beyond chips: US intelligence reportedly found extra modems on Chinese-made port cranes, and the US has banned Chinese components from connected auto systems.
  • Industrial Policy and Economic Security — Industrial policy justified on national security grounds requires weighing how much economic inefficiency is worth tolerating, since costs incurred today make a country poorer and therefore less able to project power later. The calculation is difficult because security gains are probabilistic and long-term economic costs are speculative.