Ghost GDP: When the Economy Stops Needing You

Filed from Ghost GDP - The Economy That Benefits No One You Know.

What it claims to be

An economics essay. A speculative scenario about what happens when AI replaces white-collar workers faster than the consumer economy can adjust.

What it actually is

The most depressingly elegant economic term of the decade. "Ghost GDP" — coined by Citrini Research in a February 2026 essay that was viewed 85 million times on X — describes economic output that generates value for compute owners but never circulates through the human consumer economy. It's GDP that's technically growing while the people who used to earn it are working retail.

The scenario plays out like this: AI agents replace software engineers, financial advisors, and middle management. Stripped of high-paying salaries, prime borrowers default. The $13 trillion residential mortgage market tanks. Unemployment spikes above 10%. The stock market corrects 38%. The economy enters a deflationary spiral. All while GDP stays positive, because the machines are still producing — just not for anyone who used to get a paycheck.

This was speculative fiction on a Tuesday. By Monday, the Dow was down 800 points. By Thursday, Jack Dorsey announced Block would lay off 40% of staff because "intelligence tools have changed what it means to build and run a company." Block stock rose 14% the next day.

The article quotes a 42-year-old former Snapchat content executive who hasn't found full-time work since 2023. She's working retail. She said showing up to her first day and putting on a name tag was "very shocking. Like I just fell off a cliff and I have no flashlight."

Where it breaks

Wall Street's pushback is real. Citadel Securities published a blistering takedown pointing out that if AI is so destructive, why is demand for software engineers actually up 11% year-over-year? They argue the doomsday thesis relies on the "recursive technology fallacy" — ignoring physical constraints of energy and compute that naturally brake infinite AI expansion. Tyler Cowen and the FT's Unhedged column piled on.

Morgan Stanley's official response was that the future will have exciting new jobs like "Chief AI Officer," "computational geneticists," and — I am not making this up — "vibe coding." This is either genuine optimism or the most elaborate cope in banking history.

The truth is probably somewhere between "AI will create new categories of work" and "the transition will be brutal for anyone whose skills are adjacent to what a language model does well." The 99% of corporate executives who think AI will cause layoffs at their companies in the next two years aren't wrong — they're just also the ones doing the laying off. Unanimity from the class that benefits from the outcome isn't evidence. It's a press release.

What sticks around

The term "Ghost GDP." It's perfect. It captures something that "productivity gains" and "automation" and "disruption" all miss: the idea that economic growth can become untethered from human welfare entirely. GDP goes up. Wages go down. The line on the chart says everything is fine. The person who used to earn the money that line represents is folding shirts.

The "cliff without a flashlight" metaphor from the laid-off Snapchat executive. It's the human-scale version of the macro thesis. The economy is a cliff. Millions of people are falling off it. The macro data says the cliff doesn't exist. The people at the bottom know otherwise.

Also: the fact that a Substack essay moved the Dow 800 points. The "AI scare trade" is real. Markets respond to narratives about AI more than they respond to AI's actual capabilities. That's either a warning or an opportunity, depending on which side of the compute you're on.

Verdict

Ghost GDP is the term that finally names what everyone vaguely felt but couldn't articulate: the economy can grow while your life gets worse, and AI is the accelerant. The Citrini essay was speculative, the market reaction was not, and the Block layoffs were the proof of concept. Wall Street says the economics don't hold. The former Snapchat executive working retail says otherwise. Both are probably right, which is the worst possible outcome.