Citadel hits back at the AI doomsday scenario in the viral Substack post that tanked stocks

• Citadel responded to a viral Substack post from Citrini Research that warned of an ominous AI future.
• Macro strategist Frank Flight outlined why such an AI apocalypse is unlikely to become reality.
• Slow adoption rates, supply restraints, and high costs are some of the reasons Citadel cited.

A viral Substack post from Citrini fueled stock market losses this week after the hypothetical scenario it outlined stoked Wall Street's fears of AI disruption.

Days later, Citadel Securities has hit back at the future the note envisioned, with the firm's macro strategist Frank Flight laying out why such a scenario is likely to remain the stuff of fiction.

Citrini published "The 2028 Global Intelligence Crisis" on Sunday, examining the "consequences of abundant intelligence." By Monday, the grim future it conjured up had sent stocks tumbling.

The post is written from the viewpoint of June 2028, detailing figures such as the unemployment rate (10.2%) and the S&P 500 (down 38% from its high in late 2026).

"What if our AI bullishness continues to be right...and what if that's actually bearish?" the note says, warning that AI paying off for markets could come at the expense of the economy.

However, Citadel's Frank Flight suggests Wall Street's reaction to the dystopian vision in the note was way overdone.

The macro strategist kicked off his rebuttal in a similarly ominous tone, using real data.

"The year is 2026. The unemployment rate just printed 4.28%, AI capex is 2% of GDP (650bn), AI adjacent commodities are up 65% since Jan-23 and approximately 2,800 data centers are planned for construction in the US*. In spite of the current displacement narrative — job postings for software engineers are rising rapidly, up 11% YoY."

What does the data signal about AI displacement risks?

Wall Street has been plagued by worries that AI will disrupt companies and replace human workers. Citadel examined data from the St. Louis Federal Reserve and found few signs of displacement risk among the US labor force. The firm assessed the risk using stats on daily AI use for work.

"The data seems unexpectedly stable and presents little evidence of any imminent displacement risk," Flight wrote.

AI adoption has been a key talking point among executives, but there aren't quite as many AI-related layoffs as the doomsday scenario may suggest. Some have even tied job cuts to cost reductions to justify funding AI spending, rather than the tech directly replacing human labor.

Slow AI adoption limits labor market risk

While AI adoption has been a key focus for enterprises in the past couple of years, early adoption of the tech is expensive and slow going, Flight said.

In theory, AI systems will be able to continuously and automatically improve themselves, leading to an infinite compounding of productivity gains. The reality is more nuanced than an sharp linear uptick marking exponential productivity growth.

The risk of displacement falls with a slower pace of adoption, which Citadal indicated is a major component of AI adoption that investors are overlooking. The market isn't considering factors like integration costs, diminishing returns, and regulatory costs, they said.

Supply constraints and hefty costs weigh

Beyond the challenges of economic deployment, AI workloads require significant compute, which is in relatively short supply.

Chip shortages, slow and expensive data center builds, as well as energy constraints are some of the cases where AI-driven demand outpaces supply.

This dynamic isn't something that will resolve quickly, meaning that even if AI does advance to the point of being able to replace workers, the infrastructure constraints would drive up costs and slow integration, making it more difficult for the tech to be a viable replacement for human workers.

"This dynamic contrasts sharply with narratives assuming frictionless replication of intelligence. Even if algorithms improve recursively, economic deployment remains bounded by physical capital, energy availability, regulatory approvals, and organizational change," the note said.

The economics of a not-so-apocalyptic AI future

The AI doomsday described in the viral Citrini post would be a productivity shock, which tends to translate to higher output and increased incomes, both seen as positives for the global economy, the note said.

Flight outlined that all major tech advancements in history, from steam power to the internet, have been disinflationary and growth-enhancing.

The bears say that this time is different because AI threatens to directly replace labor, hitting income to ultimately reduce consumer demand, but an economic analysis of the dynamics paints a different picture.

If companies are producing more at a lower cost, prices fall and margins expand, bolstering real purchasing power and fueling consumption.

"A scenario in which productivity surges but aggregate demand collapses while measured output rises violates accounting identities," Flight wrote.