Edward Conard

Top Ten New York Times Bestselling Author

  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “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
  • “…a must-read for serious students of economic policy.” - Glenn Hubbard, Dean, Columbia Business School, and former Chairman of the Council of Economic Advisers
  • “Unintended Consequences represents the most cogent and persuasive analysis of the Financial Crisis to date.” - Andrei Shleifer, 1999 John Bates Clark Medal Winner
  • “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
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “A full-throated defense of economic dynamism.” - The Wall Street Journal
  • “…serious thinking for serious thinkers. …a thought-provoking blueprint for growing middle- and working-class incomes.” - Mitt Romney, former Governor of Massachusetts
  • “…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
  • “Unintended Consequences should be read by anyone who takes for granted the superiority of progressive taxation and has not thought carefully about the trade-offs involved.” - The New Republic
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
Upside of Inequality Oxford Unintended Consequences

Macro Roundup


Wednesday, April 29, 2026

AI Summary. The authors estimate that exposure to negative local labor market shocks after age 50 is followed by substantial declines in cognitive test scores, with effects concentrated among men aged 51 to 64. Cognitive scores of people in the weakest-employment-growth local areas diverge from those elsewhere, reaching a one-point gap on a 28-point scale by their 70s.

Noah Arman Kouchekinia, David Neumark and Tim Bruckner National Bureau of Economic Research
Date Posted:

Kouchekinia et al find men aged 51–64 in the 1996 – 2018 waves of the Health and Retirement Study who leave employment due to exogenous shocks in local labor demand experience more rapid cognitive decline than those who continue working.

Does weak local employment growth accelerate cognitive decline after 50?

Core argument: Respondents in commuting zones in the bottom quartile of employment growth score a full point lower on the 28-point Langa-Weir cognitive scale in their 70s than those in the top quartile, despite starting their early 50s at the same level.

There are marked differences in both level and trajectory [of cognitive scores] by economic geography, displayed in Figure 3, Panel C. Respondents in the commuting zones (CZs) in the lowest quartile of employment growth start their early 50s with scores at the same level as the other three quartiles. However, a gap emerges over time, particularly as respondents enter potential retirement years. Relative to the top quartile, there is a full one-point gap [in the Langa-Weig 28 point scale] for participants in their 70s between the bottom and top quartiles of CZ employment growth. This suggests a role for work – related to local variation in employment – in slowing cognitive decline. Our main IV specification, using a Bartik shift-share instrument for employment changes, finds substantial declines over time in cognitive scores following exposure to negative labor market shocks, focusing on pre-retirement ages above 50 [Table 1 in gallery]. Our IV findings of beneficial effects of employment on forestalling cognitive decline are concentrated among men aged 51 to 64, which coheres with the notion that employment in this subgroup (compared to women or persons of retirement age) would respond most strongly to local area labor market shocks. Our evidence extends past work focusing narrowly on the retirement age window and provides further support to the notion that working to older ages may delay age-related cognitive decline.

Takeaways by Macro Roundup® AI

  1. Respondents in commuting zones in the bottom quartile of employment growth score a full point lower on the 28-point Langa-Weir cognitive scale in their 70s than those in the top quartile, despite starting their early 50s at the same level.
  2. Using a Bartik shift-share instrument for local employment changes, the authors find substantial declines over time in cognitive scores following exposure to negative labor market shocks among people above age 50.
  3. The authors' instrumental-variable findings of beneficial employment effects on forestalling cognitive decline are concentrated among men aged 51 to 64, the subgroup whose employment would be expected to respond most strongly to local labor market shocks.

Related Articles:

  • The Path to 2075 — The Positive Story of Global Aging — Extending working lives is the most effective way to offset population aging's impact on dependency ratios, and the trend is already underway: in developed economies median effective working life has risen 12% (34 to 38 years) versus a 5% gain in life expectancy, with employment's share of population up from 46.0% to 48.3%.
  • Workforce Participation for Older Americans: Warning Lights Flashing — Labor force participation for Americans 55 and older remains almost two percentage points below pre-pandemic levels, while participation for youth, prime-age women, and prime-age men has recovered or exceeded earlier levels. Participation among 55–64 year-olds in the US has been overtaken by the OECD average and Europe.
  • The Age Divide in the American Workplace — American workers over 50 earned roughly 35% more per week than workers under 30 in the late 1970s, a gap that widened to about 55% at its early-2010s peak before narrowing to 47%. Their advantage in access to top-quarter-paying management jobs also widened, from about 5 to 8.3 percentage points.

AI Summary. Despite predictions a decade ago that AI would replace radiologists, radiologists earn more than $500,000 a year and their employment continues to grow. Reading scans is a task rather than a job, and cheaper task performance can increase demand for the job.

Torsten Sløk Apollo
Date Posted:

Sløk notes an example of “Jevons paradox.” As AI diffused through radiology, employment is up ~17%, and (nominal) compensation has grown 57%. “Reading scans is a task, not a job, and when the task gets cheaper, demand for the job grows.”

Does automating a task increase demand for the workers doing it?

Core argument: Radiologists earn more than $500,000 per year and their employment continues to grow, a decade after AI was expected to replace them.

A decade ago, AI was supposed to replace radiologists. Today, radiologists make more than $500,000 per year, and their employment continues to grow. Reading scans is a task, not a job, and when the task gets cheaper, the demand for the job grows.

Takeaways by Macro Roundup® AI

  1. Radiologists earn more than $500,000 per year and their employment continues to grow, a decade after AI was expected to replace them.
  2. Reading scans is a task rather than a job, and when the task becomes cheaper, demand for the radiologist job grows.

Related Articles:

  • What The AI ‘Jobpocalypse’ Narrative Misses — Large gains in software productivity since the 1990s coincided with rising rather than falling web development employment, because demand for software grew faster than the labour needed per unit of output. Similar patterns appear in accounting, architecture, advertising, and high-tech healthcare specialisms such as lab testing and diagnostic imaging.
  • AI and the Fable of the ATMs — Bank tellers per branch fell steadily as automated teller machines spread, and total teller employment held up only because banking deregulation drove a rise in branch numbers; teller jobs are down nearly 40% from their peak. The author argues invoking this case for AI requires identifying an equivalent offsetting source of labor demand.
  • Looking for the Ladder — The article argues that the hiring downturn in occupations most exposed to generative AI reflects interest-rate sensitivity rather than AI itself, since roughly 38% of workers in the most AI-exposed quintile work in information, finance and insurance, and professional and technical services, versus under 2% in the least-exposed quintile.

AI Summary. Iran pumps about 3m barrels a day of crude plus 750,000 barrels of condensate against domestic demand near 1.9m barrels, leaving a surplus of roughly 1.85m barrels a day that a blockade is pushing into storage. Commodity-intelligence firm Kpler estimates Iran has 12–22 days of storage capacity remaining.

Javier Blas Bloomberg
Date Posted:

Given storage constraints, Iran will need to close wells and halve its oil production from pre-war levels in mid-May. However, Iran was able to restart production after shutdowns driven by the “maximum pressure” campaign without “permanent damage.”

How long can oil storage absorb a blockaded exporter's surplus crude?

Core argument: Iran pumped about 3m barrels a day of crude plus 750,000 barrels of condensate before the war, and with local demand near 1.9m barrels, roughly 1.85m barrels a day of surplus is flowing into storage under the blockade.

Before the war started, Iran was pumping about 3 million barrels a day of crude, plus another 750,000 barrels of a light oil called condensate. With local demand running close to 1.9 million barrels, that leaves Iran with a daily surplus of about 1.85 million barrels a day. With the blockade in place, those barrels are heading into storage. Helpfully for the Islamic Republic, it has more onshore and offshore capacity than most of its neighbours, giving it time before its tanks are full. Kpler, a commodity-intelligence firm, reckons that Iran probably has another 12 to 22 days of storage available, far more than the US administration anticipated when it started the blockade. Back in 2019-2020, during the first Trump administration, Iran had to cut production significantly but was able to restart its wells without significant problems over the next few months, eventually lifting total petroleum output in 2025 to a 46-year high.

Takeaways by Macro Roundup® AI

  1. Iran pumped about 3m barrels a day of crude plus 750,000 barrels of condensate before the war, and with local demand near 1.9m barrels, roughly 1.85m barrels a day of surplus is flowing into storage under the blockade.
  2. Kpler, a commodity-intelligence firm, reckons Iran has another 12 to 22 days of storage available, more than the US administration anticipated when it began the blockade, aided by onshore and offshore capacity exceeding most of its neighbours.
  3. Iran cut production significantly in 2019-2020 during the first Trump administration but restarted its wells without significant problems over the following months, eventually lifting total petroleum output in 2025 to a 46-year high.

Related Articles:

  • The Hormuz Blockade Is as Much About China as Iran — The author estimates Iran earned about $100m a day from crude sales before the war and roughly $175m a day afterward, so a US-enforced blockade would deal its economy a heavy blow. An earlier US maximum-pressure sanctions campaign cut Iranian crude earnings to no more than $10m a day without forcing Iran to yield.
  • How Fast Can Gulf Production Recover After Reopening? — Gulf oil production is likely to mostly recover within a few months of a full and safe reopening of the Strait, given limited reported physical damage to oil fields and the likelihood that Saudi Arabia and the United Arab Emirates deploy spare capacity. Full recovery could take several quarters.
  • You Can’t Print Molecules — Rebuilding after a disruption to a key shipping strait would require governments to finance defense, strategic reserves, domestic energy production, and infrastructure at the same time that inflation expectations are raising funding costs. The author argues capital will rotate out of asset-light sectors that thrived on open sea lanes and cheap energy into asset-heavy physical ones.

AI Summary. Chinese automakers BYD, Geely, and Great Wall Motor rank among the world's largest and have gained market share in Europe and Asia, with Chinese vehicles making up a quarter of sales in Mexico. US federal rules let Mexican residents and dual citizens drive noncompliant Chinese cars across the border, and Canada will allow imports of inexpensive Chinese electric vehicles.

Ryan Felton Wall Street Journal
Date Posted:

Chinese cars account for ~25% of sales in Mexico, and Canada will soon allow limited imports. According to a survey by Strategic Vision, a market research firm, ~30% of Americans would be open to buying a Chinese car.

Are Chinese cars already entering the US despite import restrictions?

Core argument: Chinese vehicles account for a quarter of total car sales in Mexico, and BYD, Geely, and Great Wall Motor rank among the world's biggest carmakers while gaining share in Europe and other parts of Asia.

BYD, Geely, and Great Wall Motor are now among the biggest carmakers in the world. They have been gobbling up market share in Europe and other parts of Asia. In Mexico, Chinese vehicles account for a quarter of total sales. Federal regulations allow Mexican residents and those with dual citizenship to drive their cars into the U.S., even if their vehicles aren’t compliant with relevant standards. That is giving Americans along the border a firsthand look at the Chinese competition. Soon, Canada will allow tens of thousands of inexpensive Chinese EVs to be imported. [Ohio Senator] Bernie Moreno [plans to introduce a bill that] would “hermetically seal” the U.S. from Chinese automakers. Chinese cars from Canada or Mexico couldn’t be driven into the country. American car companies couldn’t pursue joint ventures with Chinese automakers. Chinese car companies that own U.S. brands, such as Geely-controlled Volvo and Polestar, would have to divest themselves of those brands by 2030.

Takeaways by Macro Roundup® AI

  1. Chinese vehicles account for a quarter of total car sales in Mexico, and BYD, Geely, and Great Wall Motor rank among the world's biggest carmakers while gaining share in Europe and other parts of Asia.
  2. Federal regulations let Mexican residents and dual citizens drive their vehicles into the U.S. even when those cars do not comply with relevant standards, giving Americans along the border a firsthand look at Chinese models.
  3. Ohio Senator Bernie Moreno plans a bill to "hermetically seal" the U.S. from Chinese automakers by barring Chinese cars driven in from Canada or Mexico, prohibiting joint ventures, and requiring divestment of U.S. brands such as Volvo and Polestar by 2030.

Related Articles:

  • Cheap Chinese Cars Are Taking Over Roads From Brazil to South Africa — Inexpensive cars from Chinese makers such as BYD, Chery, Great Wall and SAIC are gaining share in emerging markets, reaching nearly 10% of sales in South Africa, 8% in Turkey and close to a third in Chile. Chinese automakers' market share outside China is projected to rise to 13% by 2030 from 3%.
  • Will China Take Over the Global Auto Industry? — The global auto market has historically been regional, with most cars sold in a region also built there. Overcapacity in China's internal combustion sector, expanding electric-vehicle capacity, ample state bank credit, and a weaker yuan could concentrate global auto production in China absent restrictions such as the US 100% tariff on Chinese electric vehicles.
  • Where Did All the Affordable Cars Go? — The average new car transaction price is around $50,000, with new cars under $20,000 essentially unavailable, while auto repair costs rose 15% over the past year to roughly $840 per visit. The author argues the US should stop protecting Detroit and open the market to Chinese cars.

AI Summary. Japan's net debt is 130% of GDP versus gross debt of 240%, yet its 30-year government bond yield sits at roughly the same level as Germany's despite far higher public debt. The author argues Bank of Japan suppression of long-term yields pushes the fiscal risk premium into a weak yen instead.

Robin Brooks Robin Brooks Substack
Date Posted:

As long as the BOJ is able to suppress long-term borrowing costs via QE, Japan’s fiscal distress will be reflected in a “weak and falling Yen” rather than high bond yields. Absent major debt reduction, Brooks argues, the Yen will decline further.

Does capping long-term bond yields shift fiscal risk into the currency?

Core argument: Japan's gross government debt stands at 240% of GDP against net debt of 130%, yet its 30-year government bond yield sits at roughly the same level as Germany's despite Germany's far lower debt.

Japan’s net debt is “only” 130% of GDP versus gross debt of 240%. The chart shows 30-year government bond yields on the vertical axis and gross government debt on the horizontal axis. What’s striking is that Japan’s 30-year yield is basically at the same level as that of Germany, even though its public debt is at a much higher level. If markets were free to set the 30-year yield without BoJ interference, it would be much higher and the Yen would appreciate. But that can’t be allowed because this would cause a fiscal crisis. So Japan’s high debt and the associated fiscal risk premium show up in a weak and falling Yen. This is just another way for markets to price fiscal distress. As long as the BoJ caps long-term yields, that isn’t going to change. There are ways out of this cul-de-sac. The most obvious one is for the government to sell some of its financial assets and use the proceeds to retire public debt.

Takeaways by Macro Roundup® AI

  1. Japan's gross government debt stands at 240% of GDP against net debt of 130%, yet its 30-year government bond yield sits at roughly the same level as Germany's despite Germany's far lower debt.
  2. Without Bank of Japan interference, Japan's 30-year yield would be much higher and the Yen would appreciate, an outcome that cannot be allowed because it would cause a fiscal crisis.
  3. Japan's high debt and fiscal risk premium show up in a weak and falling Yen rather than in long-term yields, and one way out is for the government to sell financial assets and use the proceeds to retire public debt.

Related Articles:

  • 2026 Japan Economic Outlook: Steady Fundamentals, Policy Risks Ahead — Japan's government debt-to-GDP ratio has continued to decline despite an ongoing primary deficit, because the government's interest payment rate (0.8%) is far below nominal GDP growth (3.9%), a gap the authors call the widest since 1980 and worth roughly -7 percentage points a year.
  • 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%.
  • Japan’s Debt Puzzle: Sovereign Wealth Fund from Borrowed Money — 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.

AI Summary. Germany plans to raise defence spending to €144.9bn, from about €120bn, including roughly €11.5bn in military aid to Ukraine, and government figures point to €188.4bn by 2030, equivalent to 3.7% of GDP and above NATO's 3.5% core military spending requirement for 2035.

Anne-Sylvaine Chassany Financial Times
Date Posted:

Germany will increase defense spending ~21% y/y from ~€120B this year to ~€145B in 2027. Berlin is planning on spending ~€188B annually by 2030, ~3.7% of GDP.

Will Germany exceed NATO's military spending target ahead of schedule?

Core argument: Germany will spend €144.9bn on defence in 2027, up from about €120bn this year, according to government figures circulated ahead of a formal announcement.

Berlin will spend €144.9bn for defence in 2027, from about €120bn this year, according to government figures circulated ahead of a formal announcement on Wednesday. The figure for next year includes about €11.5bn in military aid to Ukraine, about the same amount as this year. The country has also set up a special €500bn debt fund to modernise its infrastructure over the next decade. The twin moves have upended years of fiscal orthodoxy as Europe’s largest economy seeks to grow again. Berlin was planning to spend €188.4bn on defence by 2030, according to government officials — including €8.5bn for Ukraine in case the conflict has not ended by then. This would be equivalent to 3.7% of GDP, above the 3.5% Nato core military spending requirements by 2035.

Takeaways by Macro Roundup® AI

  1. Germany will spend €144.9bn on defence in 2027, up from about €120bn this year, according to government figures circulated ahead of a formal announcement.
  2. Berlin plans €188.4bn of defence spending by 2030, equivalent to 3.7% of GDP, above Nato's 3.5% core military spending requirement due by 2035, according to government officials.
  3. Germany's 2027 defence figure includes about €11.5bn in military aid to Ukraine, roughly the same as this year, with €8.5bn set aside for 2030 in case the conflict has not ended.

Related Articles:

  • Defense Spending: Macroeconomic Consequences and Trade-Offs — Peacetime defense spending booms—when the two-year average of defense outlays rises 1pp of GDP—are followed by sizable increases in real GDP, government consumption, investment, and the capital stock, with a short-lived rise in inflation. Imports accelerate while exports are broadly unchanged, widening current account and trade deficits.
  • The Intellectual Spoils of War? Defense RD, Productivity and International Spillovers — The authors estimate that a 10% increase in government-financed R&D, largely defense, is associated with a 4.3% rise in privately funded R&D across OECD countries, indicating "crowding in" rather than "crowding out." They also find international spillovers raise private R&D abroad, though the overall effect on US productivity growth is modest.
  • German Army Struggles to Get Gen Z Recruits ‘Ready for War’ — Germany's armed forces commissioner warned the country is not moving closer to its target of 203,000 active troops by 2031, as the military slightly shrank amid high dropout rates, with a quarter of new recruits leaving within six months. Conscientious objector applications reached 2,998, up from 200 in 2021.

Thursday, April 30, 2026

AI Summary. The authors find that startups with a larger share of tasks doable by generative AI cut employment beginning immediately after ChatGPT's release, with technically proficient founders restructuring teams toward higher-skill roles. Venture investors made smaller but more numerous deals, and new firm entry offset the decline in total startup employment.

Abhinav Gupta, Franklin Qian, Elena Simintzi and Yifan Sun Working Paper
Date Posted:
Is Database:
Database
Is Important:
Important

The release of ChatGPT was followed by rapid employment declines in startups with a large share of tasks that could be done by AI. The subsequent restructuring saw many new, smaller startups that employed a higher share of senior workers.

How does generative AI reshape startup hiring and venture investment?

Core argument: Startups with a larger share of tasks accomplishable by generative AI reduced employment sharply in the first quarter after ChatGPT's release, with no evidence of anticipation and monotonic declines thereafter, which the authors read as supporting a causal interpretation.

Using variation in task exposure across firms, industries, and markets, we show that startups with higher exposure to Gen AI [more precisely, a larger share of tasks accomplishable by AI], adjusted rapidly after the introduction of ChatGPT. A [negative] employment effect appears sharply in the first quarter following the release, with no evidence of anticipation, supporting a causal interpretation. Employment declines monotonically thereafter, through the end of the sample [Figure 1]. The reorganization of startup teams was concentrated among technically proficient founders, who adopted AI tools more aggressively and restructured their workforces around higher-skill, more flexible roles [Figure 3]. VCs in Gen AI-intensive sectors made smaller initial investments per firm but expanded the number of deals, broadening their exposure across a larger set of AI-related startups. This shift toward a more diversified, exploratory investment approach fueled a wave of new startup formation [Figure 4]. Employment among incumbent startups declined, but the surge in new firm entry muted the net effect on total startup employment. However, the share of managerial workers increased, indicating a shift in the structure of employment rather than a simple change in its level [Figure 2].

Takeaways by Macro Roundup® AI

  1. Startups with a larger share of tasks accomplishable by generative AI reduced employment sharply in the first quarter after ChatGPT's release, with no evidence of anticipation and monotonic declines thereafter, which the authors read as supporting a causal interpretation.
  2. Venture capitalists in generative AI-intensive sectors made smaller initial investments per firm but expanded their number of deals, a more diversified, exploratory approach that fueled a wave of new startup formation.
  3. Employment among incumbent startups declined while a surge in new firm entry muted the net effect on total startup employment, and the share of managerial workers rose, indicating a shift in the structure of employment rather than its level.

Related Articles:

  • Continued Rise in Business Formation — The gap between growth-oriented and value-oriented new business applications is widening, a divergence likely driven by a proliferation of startups built around artificial intelligence. Large language models are lowering the barriers to starting a business.
  • Attention (And Money) Is All You Need: Why Universities Are Struggling to Keep AI Talent — The annual compensation gap between top industry and academic artificial intelligence researchers widened from roughly $294,000 to $1.5m, coinciding with accelerating academic out-migration toward large incumbent technology firms. Researchers who move to industry cut paper production by about 63% while increasing patent filings nearly sixfold, consistent with a shift from open science toward proprietary innovation.
  • Which Entrepreneurs Boost Productivity? — The authors find that a minority of entrepreneurs are "transformative" — generating disproportionate productivity gains, scoring high on IQ tests, being well-educated, and hiring research and development workers. They argue that economies with more such entrepreneurs and R&D workers sustain higher long-run productivity growth.

AI Summary. U.S. gross domestic product rose at an inflation-adjusted 2% annual rate in the first quarter, below the 2.2% economists expected, while overall business investment grew at a 10.4% annual rate, the strongest in nearly three years. One estimate attributes about half of the quarter's growth to AI-related investment in equipment and intellectual property.

Harriet Torry Wall Street Journal
Date Posted:

Real GDP grew at a 2% annualized rate in Q1, driven by strong business investment, which increased 10.4% annualized. Even though most computer equipment is imported, AI-related investment may have driven ~half of GDP growth.

Is AI investment driving half of U.S. economic growth?

Core argument: U.S. GDP rose at a seasonally and inflation-adjusted 2% annual rate in the first quarter, below the 2.2% growth economists surveyed by The Wall Street Journal had expected, according to the Commerce Department.

The Commerce Department said U.S. gross domestic product—the value of all goods and services produced across the economy—rose at a seasonally and inflation adjusted 2% annual rate in the first quarter. Economists surveyed by The Wall Street Journal had expected GDP growth of 2.2% for the January-to-March period. The first quarter saw a strong increase in business spending on categories tied closely to AI, like equipment and intellectual property products, underscoring just how much AI has become an engine for the nation’s economy. Overall business investment increased at a 10.4% annual rate in the quarter, the strongest growth in nearly three years. “Even after accounting for the fact that most computer equipment is imported, AI investment seems like it accounted for about half of the overall GDP growth in the first quarter,” said Oliver Allen, an economist at Pantheon Macroeconomics.

Takeaways by Macro Roundup® AI

  1. U.S. GDP rose at a seasonally and inflation-adjusted 2% annual rate in the first quarter, below the 2.2% growth economists surveyed by The Wall Street Journal had expected, according to the Commerce Department.
  2. Overall business investment increased at a 10.4% annual rate in the first quarter, the strongest growth in nearly three years, with strong gains in AI-linked categories such as equipment and intellectual property products.
  3. Oliver Allen, an economist at Pantheon Macroeconomics, said that even after accounting for most computer equipment being imported, AI investment seems to have accounted for about half of first-quarter GDP growth.

Related Articles:

  • Corporate America Is Minting Money—and Not Just in Tech and Finance — Earnings per share for S&P 500 companies are estimated to grow more than 13% year over year for a sixth straight quarter, with sales growth and improving margins accompanying the gains. Share buybacks are contributing less to earnings growth than in many quarters, with the gap between earnings-per-share and net income growth just under 1 percentage point.
  • We’re Using So Much AI That Computing Firepower Is Running Out — Surging AI demand is straining computing capacity: Anthropic's Claude programming interface recorded 98.95% uptime over a 90-day span, below the 99.99% software companies typically promise, as its annualized revenue run rate climbed from $9bn to $14bn to $30bn. The company has capped how much usage customers can consume during weekday peak morning hours.
  • Taiwan’s Modern Miracle — Taiwan's economy has grown more than 12% over the past year, its fastest annual pace in nearly 50 years, as AI chip demand lifts net exports and pushes manufacturing to nearly 40% of output. Its trade surplus is flowing into outbound direct investment and foreign lending rather than US bonds.

AI Summary. The four largest US cloud providers — Amazon, Meta, Microsoft, and Google parent Alphabet — plan to spend $725bn on AI infrastructure this year, 77% more than the record $410bn spent the prior year. Google reported faster cloud growth than its rivals.

Stephen Morris, Ryan McMorrow, Hannah Murphy, Rafe Rosner-Uddin, et al. Financial Times
Date Posted:
Is Database:
Database

The four hyperscalers are set to spend ~$725B on CAPEX in 2026, ~77% more than in the previous year.

Are the largest cloud providers sharply increasing AI infrastructure spending?

Core argument: The big four hyperscalers — Amazon, Meta, Microsoft, and Alphabet — raised their combined AI infrastructure capital spending plans to $725bn this year, 77% above last year's record $410bn.

Google outshone its rivals in first-quarter earnings with faster cloud growth as the search giant and its Big Tech peers upped their AI infrastructure spending plans again to $725bn this year. The big four “hyperscalers,” which include Amazon, Meta, Microsoft, and Google parent Alphabet, are together expecting to spend 77% more in capital expenditures than a record $410bn last year.

Takeaways by Macro Roundup® AI

  1. The big four hyperscalers — Amazon, Meta, Microsoft, and Alphabet — raised their combined AI infrastructure capital spending plans to $725bn this year, 77% above last year's record $410bn.
  2. Google outshone its Big Tech rivals in first-quarter earnings with faster cloud growth, while joining peers in raising AI infrastructure spending plans.

Related Articles:

  • Chart of the Day: AI Capex Ate the Economy in Q4 — The Bureau of Economic Analysis revised quarterly US GDP growth down to 0.7% annualized, with the category capturing servers and computer chips contributing 0.66 percentage points and the rest of the economy about 0.04 points. Including data center structures and power buildout, AI-related capital spending may account for roughly 0.80 points, exceeding total growth.
  • The Macro Implications of the AI Capex Boom — Artificial intelligence capital spending is estimated to add roughly 1.4 percentage points to US growth in one year and about 1.5 points the next, comparable to business investment's contribution during the tech bubble. A material part of that estimate reflects profits at Nvidia and a few other chipmakers that are unlikely to be recycled back into the economy.
  • 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 White House opposes Anthropic's proposal to extend access to its Mythos artificial-intelligence model to roughly 70 more companies and organizations, which would raise the total with access to about 120. Officials cited security concerns about a model capable of conducting cyberattacks, and doubts that Anthropic has enough computing power to serve that many users without degrading government use.

Robert McMillan and Amrith Ramkumar Wall Street Journal
Date Posted:

The White House opposes Anthropic expanding the number of firms and organizations with access to Mythos from ~50 to ~120, citing both security concerns and worry that the new demand on Anthropic’s compute would hamper the government’s own use of Mythos.

Should access to a cyberattack-capable AI model be widened?

Core argument: Anthropic proposed granting roughly 70 additional companies and organizations access to its Mythos AI model, which would raise the total with access to about 120 entities, according to people familiar with the matter.

The White House opposes a plan from Anthropic to expand access to its powerful artificial-intelligence model Mythos, complicating the rollout of an AI tool capable of carrying out cyberattacks and sowing widespread disruptions online. Anthropic recently proposed letting roughly 70 additional companies and organizations use Mythos, which would have brought the total number of entities with access to about 120, people familiar with the matter said. Administration officials told the company they oppose the move because of concerns about security, the people said. Some White House officials also worried that Anthropic wouldn’t have access to enough computing power to serve that many more entities without hampering the government’s ability to use it effectively, one of the people said.

Takeaways by Macro Roundup® AI

  1. Anthropic proposed granting roughly 70 additional companies and organizations access to its Mythos AI model, which would raise the total with access to about 120 entities, according to people familiar with the matter.
  2. Administration officials told Anthropic they oppose the expansion because of security concerns, with Mythos described as capable of carrying out cyberattacks and sowing widespread disruptions online.
  3. Some White House officials worried Anthropic wouldn't have enough computing power to serve that many more entities without hampering the government's ability to use Mythos effectively, one of the people said.

Related Articles:

  • China’s Own Mythos Is Coming — America Must Prepare — The authors argue the US can extend its estimated seven-month lead in artificial intelligence by closing gaps in export controls: halting all AI chip exports to China including less advanced models, curbing smuggling networks, blocking cloud and remote access to controlled chips and US models, and restricting chipmaking equipment.
  • Misanthropic: On Mythos, Bad Human Behaviors and Systems Vulnerabilities — Anthropic reports that its Mythos model has detected thousands of high-severity cyber vulnerabilities, including ones found by chaining multiple obscure software weaknesses, and that these hacking skills are emergent rather than a designed capability. Mythos completed a corporate network attack simulation estimated to take an expert over 10 hours, which no other frontier model had solved.
  • Anthropic Model Scare Sparks Urgent Bessent, Powell Warning to Bank CEOs — The US Treasury and Federal Reserve convened Wall Street leaders over concerns that Anthropic's Mythos artificial intelligence model and similar systems could usher in an era of greater cyber risk. The meeting aimed to ensure banks are aware of the possible risks and are taking precautions to defend their systems.

AI Summary. The authors propose an economic deterrence strategy in which the United States and allied countries sign economic security agreements and use "avalanche decoupling"—ratcheting tariffs or quotas that phase out specific Chinese imports over time—to deter Chinese action against Taiwan. Applied first to sensitive goods such as drones and medicines, it could later extend to a broader share of trade.

Eyck Freymann Foreign Affairs
Date Posted:

A Taiwan crisis might be deterred by decoupling. Freymann advocates an initial “avalanche decoupling,” limited to sensitive products such as drones and drugs. If China crosses red lines, the US could “take the same approach to a broader share of trade.”

Can gradual trade decoupling deter China from attacking Taiwan?

Core argument: An economic deterrence strategy would show China that the United States and its allies could use a Taiwan crisis to trigger a gradual restructuring of the global economy on their own terms, pushing the burden of economic escalation back onto Xi Jinping.

An economic deterrence strategy would show China that the United States and its allies could use a Taiwan crisis to trigger a gradual restructuring of the global economy on their own terms and to their own benefit. A web of bilateral and multilateral economic security agreements would create the basis for this new international economic architecture, pushing the burden of economic escalation back onto Xi. The driving mechanism for such a restructuring would be “avalanche decoupling,” a process by which the United States and like-minded allies use ratcheting tariffs or quotas to phase out specific imports from China over time. In peacetime, the United States and its allies can use avalanche decoupling to break dependencies on China for a small number of highly sensitive products such as drones and medicines, leaving the vast majority of trade with China unchanged. Doing so now—before a crisis takes place—would establish the proof of concept. Then, if China crosses U.S. red lines in the future, the allies could take the same approach to a broader share of trade, possibly on an accelerated basis and working with a larger coalition.

Takeaways by Macro Roundup® AI

  1. An economic deterrence strategy would show China that the United States and its allies could use a Taiwan crisis to trigger a gradual restructuring of the global economy on their own terms, pushing the burden of economic escalation back onto Xi Jinping.
  2. "Avalanche decoupling" would use ratcheting tariffs or quotas to phase out specific imports from China over time, applied in peacetime to a small number of highly sensitive products such as drones and medicines while leaving most trade with China unchanged.
  3. A web of bilateral and multilateral economic security agreements would create the basis for a new international economic architecture, and if China crossed U.S. red lines, the allies could apply the same approach to a broader share of trade, possibly faster and with a larger coalition.

Related Articles:

  • US Notches One of Its Biggest Annual Trade Gaps Since 1960 — The US goods and services trade deficit reached $901.5bn for the year, one of the largest in data going back to 1960. US companies increased imports of computers and accessories by nearly $145bn from the prior year, an acceleration in demand that reflects large-scale artificial intelligence investment.
  • An Anatomy of the Great Reallocation in US Supply Chain Trade — China's share of US direct imports fell from a peak of about 22% in 2017 to roughly 16% in 2022, 13% in 2024, and about 9% by mid-2025 — matching its share when it joined the World Trade Organization. Vietnam, Mexico, and more recently Taiwan have been the main gainers.
  • How to Fight an Economic War — China earned about $3.4bn from exporting rare-earth elements and magnets in a year, while the U.S. Geological Survey estimates a 30% disruption to U.S. supply of the rare-earth element neodymium alone would cut U.S. GDP by 2.2%, over $600bn. This asymmetry gives Chinese export controls their leverage.