A speculative research note published on Monday has injected a sharper edge into the AI debate — and investors have taken notice.
Citrini Research’s blog post, “The 2028 Global Intelligence Crisis,” circulated widely in US markets overnight. Framed as a “thought exercise in financial history, from the future,” it sketches a hypothetical macro scenario in which rapid advances in artificial intelligence first drive extraordinary corporate margins — and later destabilise consumption, credit and the broader economy.
It was not offered as a base-case forecast. But in markets already navigating renewed US tariff uncertainty, geopolitical tensions and signs of fatigue in crowded AI trades, its themes landed heavily.
A stress test, not a forecast
The authors frame the post as preparation for potential left-tail outcomes as AI makes the economy “increasingly weird,” rather than a straight-line call on recession or collapse.
At the heart of the piece is what Citrini calls an “intelligence displacement spiral”. The scenario lays out a negative feedback loop in which AI-driven productivity gains reduce white-collar employment, boost margins and push companies to reinvest savings back into automation — further displacing workers and dampening wage growth.
The memo envisions a 2028 world in which “AI capabilities improved, companies needed fewer workers, white collar layoffs increased, displaced workers spent less, margin pressure pushed firms to invest more in AI, AI capabilities improved…” — a “negative feedback look with no natural brake”.
Phantom output, real pain
In Citrini’s stylised timeline, equity markets initially celebrate. By late 2026, AI-driven efficiencies have lifted margins significantly, pushing major indices to fresh highs. But beneath that surface strength, the consumer economy begins to erode.
Citrini highlights that the top 20% of earners account for around 65% of discretionary spending. If higher-income knowledge workers are disproportionately affected by automation, the impact on demand would exceed what headline GDP figures suggest. The report labels the result “Ghost GDP” — economic output that “shows up in the national accounts but never circulates through the real economy”.
This framing challenges the prevailing AI narrative that focuses on enhanced efficiency and margin expansion. It introduces the idea that productivity gains can coexist with weakening household demand — at least for a time.
Frictionless transactions and collapsing moats
The scenario also interrogates business models built on traditional intermediation and “human friction”.
Citrini argues that agentic AI could eliminate many of the “rent-extraction layers” underpinning sectors such as travel booking, insurance renewals and subscription services. On payments rails, once AI agents control transactions, they are assumed to optimise away costs such as card interchange fees.
In this construct, machines optimise for price and efficiency rather than brand loyalty or habit, potentially compressing margins in sectors long supported by embedded fees and network effects.
Markets react
That line of thinking fed directly into price action overnight, with both tech and financial markets hit hard globally.
US payments giants Visa, Mastercard and American Express fell between 4% and 8%, while major banks including JPMorgan and Morgan Stanley were also sharply lower and software stocks extended recent weakness.
The report was not the only driver. Markets were already contending with renewed tariff uncertainty, evolving US trade rhetoric and broader geopolitical tensions. A Financial Times report that Moody’s had warned of gaps in US accounting rules around AI-related data centre liabilities added to unease about how aggressively the AI buildout is being funded.
Updates from Anthropic relating to its Claude tools also rattled parts of the legacy software ecosystem. IBM suffered its steepest one-day fall in decades after investors focused on the potential for AI-assisted code modernisation to challenge elements of its mainframe business, while cybersecurity names were also under pressure. In some respects, the reaction echoed the very dynamic Citrini describes — AI tools not just enhancing productivity, but beginning to encroach on established revenue pools within the technology stack itself.
For investors in Australia, where the anxious tech mood has carried over on Tuesday and helped keep growth-oriented software and fintech names under pressure, the episode underscores how tightly ASX tech valuations remain linked to global AI narratives.
Many Australian-listed software companies trade on forward revenue multiples and recurring revenue assumptions similar to their US peers. When those assumptions are questioned — even hypothetically — multiples can compress quickly.
A shifting AI narrative
For the past two years, the dominant market question has been how much artificial intelligence can expand margins and accelerate earnings growth.
Citrini’s scenario shifts the lens toward distributional and second-order effects: who loses income as automation scales, how that feeds into consumption, and whether parts of the financial system built on steady cash flow assumptions become more fragile.
That does not invalidate the productivity story. But the market reaction suggests investors are increasingly willing to interrogate both sides of the equation.
For now, “The 2028 Global Intelligence Crisis” remains what its authors describe — a thought exercise. Its rapid transmission through global markets, however, shows how sensitive tech valuations have become to even hypothetical shifts in the AI narrative.
As artificial intelligence moves from promise to structural force, investors are beginning to price not just the upside — but the unintended consequences.