A thought experiment published by Citrini Research has become an unlikely market catalyst, prompting a sharp debate on Wall Street about whether artificial intelligence could accelerate growth while simultaneously damaging the broader economy.
The report, published on Substack on 22 February and titled The 2028 Global Intelligence Crisis, is presented as a scenario rather than a forecast. It nevertheless gained rapid attention among investors and was widely discussed after coverage by Gizmodo and financial media.
The core premise is simple but unsettling. Citrini and co-author Alap Shah ask what happens if bullish assumptions about AI prove broadly correct, but the economic consequences turn out to be negative for employment and consumption. In the fictional timeline, set in June 2028, the US unemployment rate reaches 10.2 per cent after a wave of white-collar displacement begins in early 2026. At the same time, productivity and headline GDP remain strong because AI agents continue to generate output at scale.
Amazon, UPS and Nike lead new wave of US layoffs linked to AI and cost cuts
This is where the report introduces the idea of “ghost GDP” — economic output that appears in national accounts but does not translate into income circulating through households and consumer markets. In the scenario, firms improve margins by replacing labour with AI systems, but the workers who lose income reduce spending. That, in turn, creates pressure for further cost cutting and automation, reinforcing the cycle. Citrini frames this as a negative feedback loop in which AI-driven productivity gains coexist with a weaker human-centred economy.
The report’s argument extends beyond software. It suggests that once AI agents become sufficiently capable, they could begin to erode business models built on intermediation or transaction friction. The text points to software-as-a-service contracts as an early pressure point, where companies may decide to build internal alternatives using agentic coding tools rather than renew expensive subscriptions. It then imagines similar disruption in delivery, payments and financial services, including scenarios in which agentic commerce and stablecoins reduce transaction costs and undermine fee-based models.
Although the publication is explicitly labelled “a scenario, not a prediction”, the market reaction indicates how sensitive investors have become to AI-related narratives. Gizmodo reported that the essay contributed to fresh selling pressure in software stocks and drew in shares linked to delivery platforms and payment processors. A Wall Street Journal report also described a broader sell-off led by technology names, with the Dow falling sharply as investors weighed the possibility of faster-than-expected disruption.
Part of the reaction appears to reflect existing market conditions rather than the report alone. By the time Citrini’s essay circulated, investors were already reassessing valuations in parts of the software sector amid concern that generative AI and agent-based tools could compress pricing power and reduce demand for some categories of enterprise software. In that environment, a detailed fictional macro scenario gave form to fears that were already present.
At the same time, the report has also drawn criticism and pushback, including from technology professionals commenting on the Substack post itself. Some argue that the scenario overstates the immediate capabilities of AI in software engineering and assumes an implausibly smooth substitution of human labour. Others contend that even if AI lowers the cost of building tools, this could expand software usage overall by making more applications economically viable. In that reading, disruption in one part of the market may be offset by new demand elsewhere.
That leaves markets facing two competing risks. One is that AI develops fast enough to disrupt employment, margins and financial assumptions more quickly than policymakers and businesses can adapt. The other is that expectations outrun reality, leading investors to misprice both threats and opportunities. Citrini’s report does not settle that question, and it does not claim to. What it has done is highlight a growing tension in current market thinking: AI is still widely viewed as a source of productivity and profit, yet it is increasingly also being treated as a source of systemic uncertainty.
For Wall Street, the episode is less about whether the 2028 scenario will happen in the form described, and more about what it reveals now. Investors are no longer debating only whether AI will transform industries. They are also debating who captures the gains, how quickly labour markets adjust, and whether headline growth can remain strong while underlying demand weakens. That is a more complex question than a standard technology rally — and one the market is now starting to price in.
