Workers Are Not Going to Be OK (Azeem Azhar, NYT)
| Author | Azeem Azhar (The New York Times, guest essay) |
|---|---|
| Published | November 17, 2026 |
| Format | Article |
| Original | Read the original |
The AI investment boom might not bust, because AI competes for the wage bill rather than the IT budget; if it pays for itself that way, the cost lands on workers through a slow, quiet squeeze, as in Engels’ pause.
Workers Are Not Going to Be OK (Azeem Azhar, NYT)
Key claims
- The setup: Anthropic is about to go public in what could be the largest IPO on record, while Wall Street worries that AI capex cannot earn its keep. US tech titans might spend $5.3 trillion between 2025 and 2030, which could require $2 trillion a year in revenue by decade’s end, 10x what the industry earns today and more than global software spend.
- The bust case assumes AI is sold like 60 years of prior software, as tools billed to IT. Azhar argues AI instead competes for the largest corporate expense, the wage bill. US white-collar wages are roughly $8-10 trillion a year. Microsoft offering buyouts to 60,000 employees to move money into AI is cited as an example of that bet.
- The leading indicator is AI spend leaving IT budgets. A survey of 1,300 executives at large companies found AI budgets tripled in a year to 3.3% of revenue, with only 15 cents of each AI dollar in IT; the rest sits in marketing, product and sales, the departments that hire.
- Another survey: 16% of American bosses say AI has helped them sell more, 60% expect it to by 2029, and 48% expect AI to reduce headcount in the next three years versus 14% expecting hiring to rise. Executives Azhar talks to say payoff is slower and harder than expected, but none plan to slow down.
- An Anthropic economics study scenario puts AI’s addition to US GDP at $2.8 trillion by 2030, enough to pay for the buildout. In the same scenario knowledge workers are nearly half again as likely to be unemployed, and labor’s share of national income falls by about as much as it did over the 50 years after 1970, possibly in under five years.
- Historical precedent: Engels’ pause, roughly 50 years of British history from 1790 in which output per worker climbed but wages barely moved. Friedrich Engels documented it from his family’s Manchester mill in 1842. A Netherlands study of automation from 2000 to 2016 found workers at automating firms earned less and struggled more to find new jobs.
- New jobs will come (about three-fifths of US workers hold job types created since 1940; Azhar was a webmaster in 1994), but possibly later, fewer, and far from the people displaced. John Maynard Keynes called technological unemployment a “temporary phase of maladjustment” in 1930, which is right only in the long run.
Notable quotes
A.I. is different. It can increasingly do more of the work previously helmed by analysts, accountants and attorneys. It competes for the largest corporate expense of all: the wage bill.
The economy might avoid a burst bubble. The cost of doing so would fall to workers, who would have to absorb the disruption through slower hiring, tepid wage growth and, ultimately, lost jobs.
The extraordinary finding: Only 15 cents of each dollar spent on A.I. lie in the I.T. department.
There was no bust. Mill owners made generational fortunes, but it was 50 years before wages began to catch up.
But a boom that paid for itself by drawing on America’s wage bill would be quiet. It would be job ads never posted, young graduates who can’t reach the first rung of advancement and wages that fall slowly but perceptibly for those who earn them.
If the politics of A.I. are bad today, they will get much worse tomorrow. That scenario merits preparation, and it is one Wall Street isn’t listening for.
How it connects
- Azeem Azhar: the author; also appears in Noah Smith and Azeem Azhar Discuss the AI Boom (podcast).
- Technological Unemployment and John Maynard Keynes: Azhar’s frame is Keynes’s 1930 warning, with the emphasis on how long the “temporary” phase lasts.
- Bubbles: an argument that the AI capex boom avoids a bust by reallocating the wage bill, not by growing a new market.
- Pairs with The Search for the AI-Native Firm (David Peterson, Angular Ventures), which makes the same “sell the work, not the software” observation from the venture side.