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
  • “…a fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “…a comprehensive explanation of the modern economy.” - Julian Robertson, Founder, Tiger Management
  • “Unintended Consequences offers deep and well-argued analyses on almost every issue.” - The New York Times
  • “…challenges misconceptions that distort our economic debates.” - Arthur Brooks, President of the American Enterprise Institute
  • “Unintended Consequences is far smarter and more thought-provoking than most economics written for the general public” - Greg Mankiw, Harvard University, Former Chairman of the Council of Economic Advisors
  • “…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 fresh argument for the productive value of inequality.” - David Autor, Professor of Economics, Massachusetts Institute of Technology
  • “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
  • “…a very valuable contribution.” - Larry Summers, former Secretary of the Treasury and director of the National Economic Council, president emeritus, Harvard University
Upside of Inequality Oxford Unintended Consequences
Buy the Books
  • Macro Roundup
  • About Roundup
  • About Ed Conard
  • Highlights
  • Topics
  • Subscribe
Edward Conard
  • twitter
  • facebook
  • linkedin
  • youtube
  • Email
  • Text Message (SMS)
  • Twitter/X
  • LinkedIn
  • Facebook
  • WhatsApp Message
Subscribe to Macro Roundup Emails
  • Mentions 892
  • Primary focus 490
Showing 490 database articles primarily about Productivity
Currently filtering by:
  • Remove Productivity
  • Remove "primary topics only" restriction
  • Remove 'Database'
Show all 7,294 articles
For whatever topics you select (currently: "Productivity", "Cronyism", "Incentives/Risk-Taking", "Innovation/Research", "Institutional Capabilities", "Intangibles", "Investment", "Startups", "Workforce Reorganization"):
Choose search scope

Your importance filter 'Database' shows fewer articles.

Remove filters to see full article counts

Innovation Can’t Be Forced, but It Can Be Quashed

AI Summary. Innovation is a gradual, collaborative process that cannot be forced but can be stifled by regulatory and institutional barriers, and it has slowed in many sectors while the lightly regulated digital sector thrives. Clearing a medical device through U.S. regulation takes 21 months versus 70 months in Germany.

Matt Ridley Wall Street Journal
Date Posted:
May 20, 2020
Is Database:
Database

Innovation is a gradual, collaborative process that can’t be forced but can be stifled by regulatory barriers. Reducing hurdles & rethinking incentives like patents can foster innovation, as seen in historical examples like the Longitude Prize.

Innovation is a gradual, collaborative process that can’t be forced but can be stifled by regulatory barriers....

Does regulation slow innovation in most sectors while lightly regulated digital thrives?

Core argument: A medical device takes 21 months to clear U.S. regulations compared with 70 months in Germany, an example of regulatory delay as an obstacle to innovation.

Innovation is a gradual, collaborative process that cannot be forced but can be stifled by regulatory and institutional barriers. Despite the perception of rapid technological advancement, innovation in many sectors has slowed, with regulatory delays being a significant obstacle. For instance, a medical device takes 21 months to clear U.S. regulations, compared to 70 months in Germany. The digital sector thrives due to lighter regulation, while other areas like nuclear power and pharmaceuticals face stagnation. The Covid-19 pandemic highlighted the need for innovation, yet vaccine development remains slow and costly, with less than a 10% success rate. Encouraging innovation requires reducing regulatory hurdles and rethinking incentives, such as using prizes instead of patents. Historical examples, like the Longitude Prize, show that unexpected solutions often arise from diverse fields. To foster innovation, policymakers should focus on creating an environment that encourages experimentation and reduces barriers to entry.

Takeaways by Macro Roundup® AI

  1. A medical device takes 21 months to clear U.S. regulations compared with 70 months in Germany, an example of regulatory delay as an obstacle to innovation.
  2. Innovation has slowed in sectors such as nuclear power and pharmaceuticals while the digital sector thrives under lighter regulation, and vaccine development remains slow and costly with a success rate below 10%.
  3. Encouraging innovation requires reducing regulatory hurdles and rethinking incentives, such as prizes instead of patents, with the Longitude Prize showing that unexpected solutions often arise from diverse fields.

“…Far from sudden “eureka” moments, innovation is nearly always a gradual process that proceeds by incremental steps. It is a team sport: There is no individual who can be credited with inventing the computer or the internet. The introduction of short-strawed wheat that revolutionized Asian agriculture in the 1960s, resulting in the defeat of famine, was driven by Manzoor Bajwa in Pakistan and M.S. Swaminathan in India. But they got the idea from Norman Borlaug in Mexico, who got it from Burton Bayles, who got it from Orville Vogel, who got it from Cecil Salmon, who got it from Gonjirô Inazuka in Japan, who got it from…at this point the trail runs cold. Once started, innovation is so inexorable it looks inevitable: 21 different people had the idea of the lightbulb independently. Many people besides the Google founders came up with search engines in the 1990s. Yet, paradoxically, innovation is impossible to predict. Almost nobody foresaw the usefulness and profitability of search engines. The first half of the 20th century brought dramatic changes in transport, the second half in communication. The switch was not the result of policy: Innovation cannot easily be forced…”

Matt Ridley, "Innovation Can’t Be Forced, but It Can Be Quashed,"Wall Street Journal, May 15, 2020, https://www.wsj.com/articles/innovation-cant-be-forced-but-it-can-be-quashed-11589555004

Innovation Can’t Be Forced, but It Can Be Quashed

The Covid-19 pandemic reveals that far from living in an age of incessant technological change, we have been neglecting innovation in exactly the areas where we most need it. Faced with a 17th-century plague, we are left to fall back mainly on the 17th-century response of quarantine and closing the theaters.

It is commonplace today to say that innovation is speeding up, but like much conventional wisdom, it is wrong. Some innovation is speeding up, certainly, but some is slowing down. Take speed itself. In my lifetime of more than sixty years, I have seen little or no improvement in the average speed of travel. Congestion on the roads and at airports has in many cases increased the scheduled travel time between two points. A modern airliner, with its high-bypass engines and less-swept wings, is designed to save fuel by going more slowly than a Boeing 707 did in the 1960s. The record for the fastest manned plane, 4,520 miles an hour, was set by the X-15 rocket plane in 1967 and remains unbroken. Boeing 747s are still flying half a century after they were launched. Concorde, the only supersonic passenger plane, is history.

Moreover, recent decades have seen innovation stalled or rejected in a number of technologies. Nuclear power has been unable to roll out plans for new reactor designs. Genetic modification of crops was effectively rejected by Europe. The flow of new pharmaceutical drugs has slowed to a trickle. Ride-sharing apps have been banned in many cities. As the investor Peter Thiel has pointed out, innovation is now largely a digital phenomenon, because bits are lightly regulated and atoms heavily regulated. On all sides we hear arguments that innovation threatens jobs, the environment, privacy and democracy.

Of immediate relevance to the current emergency, the development of vaccines has languished in the 21st century as an orphan technology, insufficiently encouraged by governments and ignored by the private sector. New vaccines are rarely profitable. By the time a company develops one for a new epidemic, the worst may be over. Last year Wayne Koff, president of the Human Vaccines Project, warned that the world was poorly prepared for a pandemic because vaccine development “is an expensive, slow and laborious process, costing billions of dollars, taking decades, with less than a 10% rate of success.”

It is not just vaccines. Throughout the economy, with the exception of the digital industry, the West is experiencing an innovation famine. The Austrian economist Joseph Schumpeter’s “perennial gale of creative destruction” has been replaced by the gentle breezes of rent-seeking. Two recent books argue that big companies in cozy cahoots with big government increasingly shy away from change, sheltered against competition by regulation and intellectual property rights. In “The Captured Economy” (2017), Brink Lindsey and Steven M. Teles make the case that to the extent that incomes have been stagnating and opportunities for social mobility drying up, the cause is not too much innovation but too little. In “The Innovation Illusion” (2016), Fredrik Erixon and Bjorn Weigel argue that Western economies have “developed a near obsession with precautions that simply cannot be married to a culture of experimentation.”

Innovation relies upon freedom to experiment and try new things, which requires sensible regulation that is permissive, encouraging and quick to give decisions. By far the surest way to rediscover rapid economic growth when the pandemic is over will be to study the regulatory delays and hurdles that have now been hastily swept aside to help innovators in medical devices and therapies, and to see whether such reforms could be applied to other parts of the economy too.

A big obstacle to innovation is the slow pace of regulatory licensing. One 2012 study found that a medical device takes around 21 months to get through the regulatory process in the U.S., from application to market, and 70 months in Germany. In the specific case of the Stratos implantable pacemaker, it took 14 months in the U.S., 40 months in France and 70 months in Italy.

There is nothing new about resistance to innovation. In 1662, William Petty, a pioneering economist, medical professor and land speculator, lamented that “the poor inventor runs the gauntloop of all petulant wits,” with everybody trying to stop him from succeeding. At the time, King Charles II, lobbied by the brewing industry, was trying to ban a novel drink called coffee because it caused fake news. “As for coffee, tea and chocolate,” the king pronounced, “I know no good they do; only the places where they are sold are convenient for persons to meet in, sit half day and discourse with all companies that come in of State matters, talking of news and broaching of lies.”

Hansom cab operators in London furiously denounced the introduction of the umbrella for offering shelter to rain-soaked pedestrians. Margarine was subjected to a decadeslong smear campaign from the dairy industry. Musicians’ unions delayed the playing of recorded music on the radio. The Horse Association of America for many years fought a rear-guard action against the tractor, and the natural-ice harvesting industry frightened people with scares about the safety of refrigerators. You may laugh, but the same happens today: Incumbent vested interests, overcautious regulators, opportunistic activists and rent-seeking patent holders combine to oppose or delay almost every innovation.

There is no good evidence that patents are helpful, let alone necessary, in encouraging innovation.

Surprisingly, there is no good evidence that patents are helpful, let alone necessary, in encouraging innovation. A 2002 study by Josh Lerner, an economist at Harvard Business School, looked at 177 cases of strengthened patent policy in 60 countries over more than a century, finding that “these policy changes did not spur innovation.” James Watt, Samuel Morse, Guglielmo Marconi, the Wright brothers and many others wasted the best years of their lives in court defending their intellectual property, when they might have been busy developing new devices.

The expiration of patents often results in a burst of innovation, as with 3-D printing, where the recent lapse of three key patents has resulted in notable improvements in quality and a drop in price. The historian Anton Howes, of the Royal Society of Arts in London, points out that the French government bought out Louis Daguerre’s patent for photography in 1839 and made the technology freely available, unleashing a burst of creative innovation. Dr. Howes argues, “As we look to fight coronavirus and any future pandemics, we should perhaps consider which patents—for antivirals, vaccines, ventilators and other hygienic equipment—might be bought out in order to remove…innovation bottlenecks.”

What is innovation and how is it best encouraged? Policy makers tend to assume it is the result of clever people having clever thoughts in ivory towers. But science is just as often the fruit as the seed of technology, and invention is different from innovation. To illustrate the difference, Charles Townes, who won the Nobel Prize for the physics behind the laser in 1964, was fond of telling the story of a beaver and a rabbit looking up at the Hoover Dam. “No, I didn’t build it myself,” says the beaver. “But it’s based on an idea of mine.” Discoverers and inventors often feel they get too little credit or profit from a good idea, forgetting just how much effort was needed to turn that idea into a workable, affordable innovation that actually delivered benefits to people.

Thomas Edison understood better than anybody that trial and error is the key to turning an invention into a useful innovation. In developing the nickel-iron battery, his employees undertook 50,000 experiments. Developing a new technology, he famously said, is 1% inspiration and 99% perspiration. Jeff Bezos makes the same point: “Being wrong might hurt you a bit, but being slow will kill you. If you can increase the number of experiments you try from a hundred to a thousand, you dramatically increase the number of innovations you produce.” It turns out that continuous tinkering to develop and refine a better product is much more important than protecting what you’ve already created.

Once started, innovation is so inexorable it looks inevitable, yet it’s nearly impossible to predict.

Far from sudden “eureka” moments, innovation is nearly always a gradual process that proceeds by incremental steps. It is a team sport: There is no individual who can be credited with inventing the computer or the internet. The introduction of short-strawed wheat that revolutionized Asian agriculture in the 1960s, resulting in the defeat of famine, was driven by Manzoor Bajwa in Pakistan and M.S. Swaminathan in India. But they got the idea from Norman Borlaug in Mexico, who got it from Burton Bayles, who got it from Orville Vogel, who got it from Cecil Salmon, who got it from Gonjirô Inazuka in Japan, who got it from…at this point the trail runs cold.

Once started, innovation is so inexorable it looks inevitable: 21 different people had the idea of the lightbulb independently. Many people besides the Google founders came up with search engines in the 1990s. Yet, paradoxically, innovation is impossible to predict. Almost nobody foresaw the usefulness and profitability of search engines. The first half of the 20th century brought dramatic changes in transport, the second half in communication. The switch was not the result of policy: Innovation cannot easily be forced.

Given this unpredictability, politicians should rethink the incentives for innovation. One option is to expand the use of prizes, to replace reliance on grants, subsidies and patents. Britain’s famous Longitude Prize, offered in 1714 for accurately measuring longitude at sea, elicited a solution from an unexpected direction: accurate and robust clocks made by a humble clockmaker, John Harrison. Similar serendipity happens today. One study of the online problem-sharing forum known as Innocentive, where organizations can reward crowdsourced solutions to problems that baffle them, found that “the further the focal problem was from the solvers’ field of expertise, the more likely they were to solve it.” Teflon, Kevlar and the Post-it Note are all examples of useful things developed by people looking for something completely different.

The Nobel Prize-winning economist Michael Kremer came up with a concept called the Advance Market Commitment that would fine-tune prizes as incentives to innovate. In 2007, the Gates Foundation committed $1.5 billion to a prize fund to find a vaccine for Pneumococcus bacteria for use in developing countries, where no pharmaceutical company could make money from it. But rather than just winning a lump sum for meeting the goal, companies were assured that if they succeeded, they would win a contract that paid out for 10 years at a good price. The prize money effectively topped up the sum received by the pharmaceutical firm for every vaccine sold. The result of the auction was three good vaccines, costing $2 per dose, which have been given to 150 million children, saving 700,000 lives.

The world would benefit from more innovation, and it need not be high-tech. Consider just one remarkable development that has saved millions of lives from malaria in recent decades: the insecticide-treated mosquito net.

In June 1983, in the West African country of Burkina Faso, a group of French and Vietnamese scientists together with African colleagues began an experiment. They installed mosquito nets in 36 huts specially equipped with mosquito traps. Half the nets were treated with the insecticide permethrin and half had holes torn in them to simulate the wear and tear of everyday life. For five months, volunteers slept in the huts, while every mosquito that entered or tried to leave the huts was collected, dead or alive. The results showed that the presence of a permethrin-treated net, whether intact or torn, repelled mosquitoes, reducing the number entering the huts by about 70% and increasing the rate at which the insects left the hut from 25% to 97%.

This simple, careful experiment proved to be a breakthrough in the control of malaria in Africa. Supported by the Gates Foundation, impregnated bednets first started to be used on a wide scale in 2003, and that very year malaria mortality began to decline. According to a recent study in the journal Nature, insecticide-treated mosquito nets account for 70% of the six million lives saved world-wide from malaria in recent years, twice as high a percentage as antimalarial drugs and insecticide sprays put together. Globally, the death rate from malaria has halved since 2003.

Dealing with Covid-19 has forcibly reminded governments of the value of innovation. But if we are to get faster vaccines and treatments—and better still, more innovation across all fields in the future—then innovators need to be freed from the shackles that hold them back.

  • Productivity
    • Innovation/Research
Previous articleMay 20, 2020Killer Incentives: Status Competition and Pilot Performance during World War IIAnalysis of >5k German fighter pilots reveals that awards & status competition significantly influenced performance, with disparities across skill levels. Skilled pilots increased victory scores without higher mortality rates, while average pilots showed minimal improvement in victories & higher mortality.Next articleMay 21, 2020Emulation, Inequality, and Work Hours: Was Thorsten Veblen Right?Work hours increase with income inequality, with greater inequality predicting longer work hours, according to @SamuelBowles and @YongjinPark in The Economic Journal.
Showing 489 database articles primarily about either Productivity, Cronyism, Incentives/Risk-Taking, Innovation/Research, Institutional Capabilities, Intangibles, Investment, Startups, or Workforce Reorganization

How Much Revenue Does The AI Buildout Need In 2030?

AI Summary. The authors estimate AI capital spending of $560bn in 2026 and, in a high-investment case, $3.5tn of net AI capital in service by 2030, requiring roughly $1.3tn in annual infrastructure revenue once depreciation, operating costs and a 15% return are covered. That implies 7.4x growth from current annualized revenues, with downstream revenues of $1.9–2.6tn.

William Gildea and Azeem Azhar Exponential View
Date Posted:
October 1, 2026
Is Database:
Database

Beginning with capex projections, Gildea/Azhar derive for their high-growth case a $1.3T annual AI-infrastructure revenue hurdle for 2030. This covers $771B of depreciation and operating costs, plus a 15% return on $3.5T of net capital in service.

How much annual revenue must AI infrastructure generate to justify investment?

Core argument: In a high investment case with $3.5tn of net AI capital in service, the authors estimate AI infrastructure must generate $1.3tn of annual revenue by 2030, requiring 7.4x growth (59% CAGR) from today’s annualized revenues.

We estimate AI CapEx to be $560B in 2026, 64% of the total declared by the hyperscalers. Unfinished facilities can’t generate revenue, so the investment bill and the productive asset base will grow at different speeds. [In the high investment case], by 2030 we see $3.5T of net AI capital in service. The largest component [of the revenue hurdle] is the $555B in compute depreciation. We model lifespans at the current 5-6 years. Shell depreciation and OpEx combined contribute about a further $216B to the required revenue hurdle. [Adding a 15% return on net capital brings required annual infrastructure revenue to $1.3T]. Reaching the $1.3T hurdle would represent 7.4x growth (59% CAGR) from today’s annualized revenues. For its customers to earn 30-50% gross margins, downstream revenues need to be $1.9-2.6T. [Editor’s note: Van Nieuwerburgh’s far larger figure of $3.7T covers mature revenues from projects completed through 2032, using commissioning capital and a 50% infrastructure operating cash-flow margin. This estimate instead charges depreciation and operating costs plus 15% on net capital in service in 2030. Different horizons and cost conventions prevent a direct comparison. Van Nieuwerburgh also has a much higher capex estimate.]

Takeaways by Macro Roundup® AI

  1. In a high investment case with $3.5tn of net AI capital in service, the authors estimate AI infrastructure must generate $1.3tn of annual revenue by 2030, requiring 7.4x growth (59% CAGR) from today’s annualized revenues.
  2. Compute depreciation of $555bn, modeled on 5-6 year lifespans, is the largest component of the 2030 revenue hurdle, with shell depreciation and operating expenses adding about $216bn and a 15% return on net capital completing the $1.3tn total.

Related Articles:

  • Financing the AI Buildout — The authors estimate roughly 57 gigawatts of operating US data center capacity against a planned development pipeline of about 509 gigawatts, which is not a completion forecast since projects may be delayed, scaled back, or abandoned. Their central scenario has 182.8 gigawatts of added capacity becoming operational through 2032, 117.2 gigawatts completed later, and 226.9 gigawatts never completed.
  • What Investment Data Implies about the AI Transition — The five largest U.S. technology firms are scaling AI capital spending about 50% a year, from $155bn to a forecast $755bn. Embedding this in a two-sector model with rare productivity booms, the authors estimate AI adds roughly 5, 20, or 58pp to cumulative growth over a seven-year transition, depending on scenario.
  • Rear Window — Major cloud providers posted annual revenue growth of 37% at Amazon Web Services, 43% at Microsoft Azure and 82% at Google Cloud, with rising revenue backlogs of $514bn and $496bn respectively at Google Cloud and Amazon Web Services. Earnings growth appears highly contingent on continued hyperscaler capital spending and research, which rose from $460bn to $870bn.
  • Investment
  • GDP
    • Financial Markets
  • Productivity
    • Innovation/Research

AI and Entrepreneurship: What is Really Happening?

AI Summary. Entrepreneurship by businesses with employees is flat or declining, including in AI-intensive information and professional services, where establishment entry has been flat or falling and young-firm employment in AI-exposed industries is falling. Applications for likely solo businesses with no employees rose steeply in those AI-intensive sectors, but self-employment as a main job is flat, suggesting side hustles.

Nathan Goldschlag Economic Innovation Group
Date Posted:
September 30, 2026
Is Database:
Database

Nathan Goldschlag finds that available federal data do not yet establish that AI is boosting entrepreneurship. “Across different cuts of the data, there is no clear evidence of AI affecting entrepreneurship, either for likely employers or nonemployers.”

Is AI driving a boom in new employer businesses?

Core argument: Business applications likely to become employer businesses were flat in 2025 and began rising in 2026, with fewer applications of the kind having the highest chance of becoming an employer business.

Employer entrepreneurship is at best rising modestly, and at worst has been falling for several years already. The number of business applications likely to become an employer business was flat in 2025 and began rising in 2026. Fewer of those applications have been the kind that have the highest chance of becoming an employer. Nonemployer entrepreneurship, or “solopreneurship,” is likely rising, but not as the main job of the solopreneurs. These are probably side hustles. Likely nonemployer business applications rose significantly in 2025 and early 2026. But the share of workers claiming to be self-employed nonemployers in their main job — in other words, business owners with no employees — is flat (though noisy). AI-intensive sectors and industries are no exception. Establishment entry in AI-intensive sectors (information and professional services) has been flat or falling in recent years.

Takeaways by Macro Roundup® AI

  1. Business applications likely to become employer businesses were flat in 2025 and began rising in 2026, with fewer applications of the kind having the highest chance of becoming an employer business.
  2. Likely nonemployer business applications rose significantly in 2025 and early 2026, going vertical in the most AI-intensive sectors of information and professional services, while other sectors kept pace until early 2026 and then flattened.
  3. The share of workers reporting self-employment as nonemployers in their main job is flat, though noisy, including among workers in AI-exposed occupations in 2025 and early 2026.

Related Articles:

  • The Age Of The Solopreneur — A surge in US business formation is being driven by solo entrepreneurs, with growth in AI tool usage and in solopreneur activity in high-AI-adoption sectors suggesting artificial intelligence accounts for a meaningful portion of the increase. The authors argue AI is lowering barriers to starting and growing a business by expanding what one-person firms can do.
  • America Has Become An Entrepreneur’s Paradise — US growth in new-business applications is concentrated in professional services (up 25% over four years), retail (18%), and health care and social assistance (18%) rather than artificial-intelligence startups. The share of applications the Census Bureau deems likely to hire at least one employee has fallen to 30% from 38%.
  • 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.
  • Startups
  • Productivity
    • Innovation/Research
    • Investment

Tech's Trillion-Dollar Internal Inconsistency

AI Summary. Wall Street analysts covering technology project the sector's operating cash flow will more than double to roughly $2.4tn by 2028, a gain of over $1.2tn, while analysts covering the rest of the S&P 500 — technology's customers — project those firms will add far less cash flow. These sector-by-sector forecasts are internally inconsistent.

Torsten Sløk Apollo
Date Posted:
September 29, 2026
Is Database:
Database

Tech analysts expect sector operating cash flow to more than double to ~$2.4T by 2028. Analysts covering other S&P 500 sectors (tech’s customers) see far less cash growth in their sectors. Slok “raises the question of who exactly will be writing all those checks to buy AI services.”

Are tech cash flow forecasts inconsistent with customers' projected growth?

Core argument: Analysts covering the tech sector expect its operating cash flow to more than double to roughly $2.4tn by 2028, an increase of over $1.2tn.

Wall Street equity analysts work in sector silos, and when you add up their forecasts, the numbers are internally inconsistent. The analysts covering tech expect the sector's operating cash flow to more than double to roughly $2.4 trillion by 2028, an increase of over $1.2 trillion. Meanwhile, the analysts covering the other sectors in the S&P 500, which are tech's customers, expect those companies to add much less operating cash flow. In other words, the tech silo is betting on a future in which demand for AI and tech services explodes, while the silos covering the companies that would pay for those services see a much more modest outlook. Both cannot be right at the same time. The bottom line is that either tech's customers will generate a lot more cash than their analysts expect, or tech's cash flow forecasts are too optimistic, which raises the question of who exactly will be writing all those checks to buy AI services.

Takeaways by Macro Roundup® AI

  1. Analysts covering the tech sector expect its operating cash flow to more than double to roughly $2.4tn by 2028, an increase of over $1.2tn.
  2. Analysts covering the other S&P 500 sectors, which are tech's customers, expect those companies to add much less operating cash flow, leaving the combined forecasts internally inconsistent.
  3. Either tech's customers will generate much more cash than their analysts expect, or tech's cash flow forecasts are too optimistic, which raises the question of who will pay for AI services.

Related Articles:

  • A Catalyst For The AI Bubble Break — Historically, controlling for valuations, a 1% rise in the share of the market represented by initial public offerings has been associated with roughly a 4% decline in prices. With large artificial-intelligence company listings expected and share issuance exceeding buybacks, equity supply could shift from shrinking about 1% a year to growing near 5%, implying returns roughly 20% below average.
  • Financing the AI Buildout — The authors estimate roughly 57 gigawatts of operating US data center capacity against a planned development pipeline of about 509 gigawatts, which is not a completion forecast since projects may be delayed, scaled back, or abandoned. Their central scenario has 182.8 gigawatts of added capacity becoming operational through 2032, 117.2 gigawatts completed later, and 226.9 gigawatts never completed.
  • Rear Window — Major cloud providers posted annual revenue growth of 37% at Amazon Web Services, 43% at Microsoft Azure and 82% at Google Cloud, with rising revenue backlogs of $514bn and $496bn respectively at Google Cloud and Amazon Web Services. Earnings growth appears highly contingent on continued hyperscaler capital spending and research, which rose from $460bn to $870bn.
  • Investment
  • GDP
    • Financial Markets
  • Productivity
    • Innovation/Research

Nine Percent of GDP

AI Summary. An analysis estimates that for US data center owners to break even on AI investments, the sector would need about $3.725tn in annual compute revenue by 2032 — roughly 9.2% of projected 2032 GDP. The estimate assumes 183 GW of US capacity completed, a $9.61tn capital base, a 10% discount rate, and a 50% operating cash flow margin.

Hanno Lustig The Two Cents
Date Posted:
September 28, 2026
Is Database:
Database

The US AI-buildout is projected to average ~3.6% of GDP over 2025–32. Lustig notes data-center operators will need ~$3.725T in annual compute revenue by 2032, 9.2% of projected GDP, to earn a 10% unlevered return on the buildout.

How much annual revenue must AI data centers generate to break even?

Core argument: Stijn’s baseline estimate holds that data center owners would need $3.725 trillion of annual revenue by 2032 — roughly 9.2% of projected 2032 GDP — to break even on AI investment at a 10% discount rate.

What do you have to believe about future revenue in order for all data center owners to to break even on their AI investment? Using a 10% discount rate, Stijn’s baseline number is 3.725 trillion of annual revenue by 2032, which would amount to roughly 9.2% of projected 2032 GDP. This number is the top line earned from selling compute in aggregate by whoever operates the 182.7 GW of US capacity completed by 2032.2. The assumptions behind that number: 183 GW of U.S. data-center capacity completed between 2025 and 2032. $8.54 trillion of construction cost, grossed up to a $9.61 trillion capital base to account for the return on capital tied up during construction. A 10% discount rate. A 6-year life for IT equipment (68% of cost) and a 20-year life for everything else (32%). A 50% operating cash flow margin.

Takeaways by Macro Roundup® AI

  1. Stijn’s baseline estimate holds that data center owners would need $3.725 trillion of annual revenue by 2032 — roughly 9.2% of projected 2032 GDP — to break even on AI investment at a 10% discount rate.
  2. The break-even calculation assumes 183 GW of U.S. data-center capacity completed between 2025 and 2032 at $8.54 trillion of construction cost, grossed up to a $9.61 trillion capital base for return on capital tied up during construction.
  3. The estimate assumes a 6-year life for IT equipment, which accounts for 68% of cost, a 20-year life for the remaining 32%, and a 50% operating cash flow margin.

Related Articles:

  • Financing the AI Buildout — The authors estimate roughly 57 gigawatts of operating US data center capacity against a planned development pipeline of about 509 gigawatts, which is not a completion forecast since projects may be delayed, scaled back, or abandoned. Their central scenario has 182.8 gigawatts of added capacity becoming operational through 2032, 117.2 gigawatts completed later, and 226.9 gigawatts never completed.
  • A Catalyst For The AI Bubble Break — Historically, controlling for valuations, a 1% rise in the share of the market represented by initial public offerings has been associated with roughly a 4% decline in prices. With large artificial-intelligence company listings expected and share issuance exceeding buybacks, equity supply could shift from shrinking about 1% a year to growing near 5%, implying returns roughly 20% below average.
  • Rear Window — Major cloud providers posted annual revenue growth of 37% at Amazon Web Services, 43% at Microsoft Azure and 82% at Google Cloud, with rising revenue backlogs of $514bn and $496bn respectively at Google Cloud and Amazon Web Services. Earnings growth appears highly contingent on continued hyperscaler capital spending and research, which rose from $460bn to $870bn.
  • Investment
  • GDP
    • Financial Markets
  • Productivity
    • Innovation/Research

Financing the AI Buildout

AI Summary. The authors estimate roughly 57 gigawatts of operating US data center capacity against a planned development pipeline of about 509 gigawatts, which is not a completion forecast since projects may be delayed, scaled back, or abandoned. Their central scenario has 182.8 gigawatts of added capacity becoming operational through 2032, 117.2 gigawatts completed later, and 226.9 gigawatts never completed.

Stijn Van Nieuwerburgh Brookings Papers on Economic Activity
Date Posted:
September 24, 2026
Is Database:
Database
Is Important:
Important

Stijn Van Nieuwerburgh projects ~$10.3T will be spent on the AI buildout over 2025–2032, averaging 3.63% of GDP. He estimates 2/3 of that spending reflects IT equipment and one-third real estate and associated power infrastructure.

How much of the planned data center pipeline gets built?

Core argument: Using Cleanview project-level data from July 2026, the authors estimate roughly 57 GW of operating U.S. data center capacity against a planned development pipeline of roughly 509 GW, which they stress is not a completion forecast.

The required investment is enormous. A representative 200 MW AI training campus costs roughly $8.2 billion, about two-thirds of which reflects IT equipment and one-third the real estate and associated power infrastructure. We estimate that 183 GW is completed by 2032 with a further 118 GW completed after 2032. This pipeline will trigger about $10.3 trillion dollars of investment over 2025–2032. At an average of 3.63% of GDP per year, the projected buildout would be larger relative to the economy than the major U.S. canal, railroad, electrification, highway, and telecommunications investment booms. It would double the electricity consumption of the entire U.S. residential sector.

Takeaways by Macro Roundup® AI

  1. Using Cleanview project-level data from July 2026, the authors estimate roughly 57 GW of operating U.S. data center capacity against a planned development pipeline of roughly 509 GW, which they stress is not a completion forecast.
  2. The authors’ central scenario allocates the announced pipeline so that 182.8 GW of additional U.S. capacity becomes operational during 2025–32, including 19.0 GW tied to projects already operational in 2025 or 2026.
  3. The authors’ scenario has 117.2 GW of planned data center capacity completed after 2032 and 226.9 GW never completed, applying lower realization rates to more distant and less fully documented projects.

Related Articles:

  • Rear Window — Major cloud providers posted annual revenue growth of 37% at Amazon Web Services, 43% at Microsoft Azure and 82% at Google Cloud, with rising revenue backlogs of $514bn and $496bn respectively at Google Cloud and Amazon Web Services. Earnings growth appears highly contingent on continued hyperscaler capital spending and research, which rose from $460bn to $870bn.
  • AI Set to be Largest CapEx Cycle Ever … and Soon Majority Externally Financed — Hyperscaler and data-center investment is tracking toward $1.5tn annually by 2030, which in real terms would exceed the 2005-06 housing boom, the 1990s fiber peak, and the 1882 railroad peak combined. External debt and equity are projected to fund close to 90% of the buildout, making it dependent on credit markets.
  • The Summer I Turned Pretty — A warning sign in a boom cycle is when companies closest to final demand weaken while capital spending beneficiaries keep thriving, as when telecom operators' shares rolled over while communications equipment stocks kept rising. Semiconductor stock prices and free cash flow are soaring while large cloud providers' shares stagnate and their free cash flow plummets.
  • Investment
  • GDP
    • Financial Markets
    • Growth
  • Productivity
    • Innovation/Research

Moonshot Capitalism: AI Rewrites The Venture Capital Playbook

AI Summary. Global investment in "deep tech" — companies whose products rest on major scientific or engineering advances — has exceeded $150bn since the start of 2024, more than the $133bn invested in the decade through 2019, excluding sums poured into AI start-ups. Failed battery and electric-vehicle bets from an earlier peak illustrate the risks of such moonshot dealmaking.

Tim Bradshaw Financial Times
Date Posted:
September 10, 2026
Is Database:
Database

Since the start of 2024, more than $150B of venture capital has been invested into non-AI “deep tech” firms whose products are rooted in significant engineering advances, exceeding the $133B invested in such firms btw 2010 and 2019.

Is the AI boom pushing venture capital toward riskier moonshot bets?

Core argument: Global "deep tech" investment excluding AI start-ups has exceeded $150bn since the start of 2024, more than the $133bn invested in the entire decade through 2019, according to Dealroom.

The AI boom is fuelling a resurgence in ambitious “moonshot” bets, as early SpaceX backers’ huge returns and falling valuations for traditional software companies force tech investors to embrace riskier and more capital-intensive dealmaking. Excluding the giant sums ploughed into AI start-ups, global investment in “deep tech” — companies whose products are rooted in big scientific or engineering advances — has exceeded $150bn since the start of 2024, more than the $133bn in the entire decade to the end of 2019, according to Dealroom. This year’s deep-tech investments have not yet surpassed 2021’s peak, which was propelled by battery and electric vehicle deals for the likes of Rivian and Northvolt — many of which turned sour, highlighting the risks involved in moonshot dealmaking.

Takeaways by Macro Roundup® AI

  1. Global "deep tech" investment excluding AI start-ups has exceeded $150bn since the start of 2024, more than the $133bn invested in the entire decade through 2019, according to Dealroom.
  2. Huge returns for early SpaceX backers and falling valuations for traditional software companies are pushing tech investors toward riskier, more capital-intensive "moonshot" dealmaking.
  3. Deep-tech investment has not yet surpassed the 2021 peak, which was propelled by battery and electric vehicle deals for companies such as Rivian and Northvolt, many of which turned sour.

Related Articles:

  • Capital Is Making a Comeback — Tangible investment in property, factories and equipment fell from 12.5% to 8.5% of private gross domestic product, while intangible investment in intellectual property, software and process knowledge rose from about 11.5% to 16.75%. The author argues demand for tangible capital investment will now have to rise.
  • Public to Private Equity in the United States: A Long-Term Look — An analysis of more than 31,000 global venture capital deals finds 62% lost money and over half lost 50–100% of invested capital, though the distribution of returns has much fatter tails than buyouts or public equities. Venture funds averaged a public market equivalent ratio of about 1.4, indicating returns above public markets.
  • Gross and Net US Investment — Net US investment has fallen from roughly 40% of gross investment in the 1970s to about 25% in the past decade, meaning about three-quarters of gross investment now just replaces depreciating capital. The author attributes this gap to modern investment being more information-technology-related, which depreciates faster and requires more frequent replacement.
  • Investment
  • GDP
    • Financial Markets
  • Productivity
    • Innovation/Research
© Copyright 2026 Coherent Research Institute · All Rights Reserved · Privacy · Terms