AI was sold as a deflation machine

AI was sold as a deflation machine. So far it is behaving like an inflation machine.

Derek Thompson makes a point that deserves more airtime. For years the builders promised "Moore's Law for Everything". In 2021 Sam Altman predicted that the price of many kinds of labour would fall towards zero, pulling down the cost of goods and services. Marc Andreessen imagined AI driving prices across the board to virtually zero. The price of that abundance, we were told, would be mass unemployment.

What has arrived is close to the opposite. Almost everybody still has a job, and things are getting more expensive. Thompson notes that the Conference Board's measure of US consumer confidence has just hit its lowest mark since 2014, while the AI buildout is, in his words, pushing up prices across the economy as it mints a generation of billionaires.

That gap between promise and lived experience explains a lot of the public hostility to AI. People were told their children's jobs might disappear in exchange for cheap everything. They are getting the anxiety without the cheap everything.

The strategic lesson is about sequencing. The costs of a general-purpose technology arrive first and land locally: a higher bill, a development application at the end of the street. The benefits arrive later and spread thinly. Political licence gets spent in that interval.

Any country building data centres at scale will run the same experiment. The industry that explains the interval honestly, rather than repeating the abundance pitch, is more likely to keep its licence to operate.

Abundance is a forecast. The bill is a fact.

Source: Derek Thompson, "We're Missing a Key Reason Why Americans Hate AI" (2 October 2026), https://lnkd.in/grBV9zmJ

#AI #Infrastructure #Strategy