A 38% stock crash and a ghost GDP: what a viral report says about the future with AI

On February 23, a new wave of panic selling of stocks was triggered by a report by a small research company Citrini Research that went viral online. Its analysts described a scenario of a possible near future in which AI has become so efficient that it has destroyed the basis of the economy - consumption based on people's income. What started out as a technological triumph led to mass layoffs of white-collar workers and, ultimately, a stock market crash.

Prominent shortstop Michael Burry, on social media, described the study as brilliant and tough. And added: "And you're still saying I'm the one with the bearish attitude?" Some analysts and technology executives found the scenario described unrealistic, Barron's wrote. Nevertheless, all of the stocks mentioned in Citrini's report fell several percent in Monday trading.

The Global Intelligence Crisis report is not a forecast, but an attempt to conduct a thought experiment and model AI-related risks that markets may be overlooking, Citrini analysts and co-author Alap Shah, chief investment officer at Lotus Technology Management, wrote. In their study, they try to describe a hypothetical world from a point in 2028 where the rapid development of AI has led to a dramatic decline in the number of office workers.

In their scenario, the AI boom is transforming the economy, but not in a positive way, as AI proponents like Elon Musk propagate. In this world, a "ghost GDP" is generated - "output that shows up in national accounts but never circulates in the real economy." This is because artificial intelligence, having become a full-fledged substitute for human labor, has not become a consumer of increased productivity.

The lack of consumer demand set off an "intelligence displacement spiral": AI capabilities improved, companies needed fewer workers, white-collar layoffs rose, those laid off spent less, and margin pressure forced firms to invest more in AI.

In parallel, the authors record the collapse of the "mediation economy". Companies making money on the complexity of actions or laziness of people (SaaS, banks, insurance, real estate) lost their advantages as AI agents started to find cheap ways to bypass intermediaries.

Ultimately, a chain of correlated bets on white-collar productivity growth led to a massive stock crash. By the summer of 2028, the S&P 500 had fallen 38% from its October 2026 highs. As analysts wrote, the markets collapsed when it became clear:

The end of 2025: an inflection point. The surge in agent AI capabilities has changed procurement logic. Companies are not renewing expensive SaaS contracts, but rather creating their own counterparts. For example, a Fortune 500 procurement manager got a 30% discount on a SaaS contract renewal because he threatened the vendor that he would replace their product with OpenAI.

Monday.com, Zapier, and Asana - representatives of the "long tail" of SaaS - are in a much worse situation. Investors expect AI to hit them hardest: although they may account for a third of the costs of a typical enterprise IT stack, they are clearly vulnerable.

2026: the euphoria before the storm. The S&P 500 approaches 8,000 points, the Nasdaq breaks through 30,000. The first mass white-collar layoffs begin. Corporate profits soar, margins expand, and record profits are funneled back into AI computing.

But in November, a report from IT company ServiceNow showed a slowdown in net new annualized contract value growth from 23% to 14%. It turned out that staff cuts were destroying ServiceNow's revenue. Shares fell 18%.

Early 2027: the collapse of mediation and the "death" of brands. AI agents have begun to manage people's consumption 24/7. The average American consumes 400,000 tokens (units of text and data processing) per day - 10 times more than a year ago. This is disrupting entire industries that made money from monetizing human limitations - lack of time, patience, or habit.

The face of the crisis was the food and goods delivery service DoorDash: the business collapsed because AI agents find more favorable prices for consumers. Tourism, insurance and real estate have also suffered: in insurance, 15-20% of premiums that companies used to earn from passive renewal of policies have disappeared; in real estate, commissions are falling below 1%.

April-September 2027: Financial contagion. AI agents conduct transactions via stablecoins on Solana or Ethereum L2 networks, bypassing bank fees. Mastercard report registers a slowdown in purchase volume growth to 3.4%. Shares of American Express, Synchrony, Capital One and Discover fell more than 10%.

In September, there was a default in private lending - Zendesk defaulted on a $5 billion debt to Blackstone, Apollo, Blue Owl and HPS. The loan was against "recurring revenue," but the standalone AI agents made it irregular.

The end of 2027: systemic risk. In November 2027, the stock market experiences a massive crash. It started when the market realized that profit growth through layoffs was a trap. Companies like Salesforce or Microsoft became more efficient, but they destroyed their core customer. When laid-off employees stop paying their Capital One credit cards and mortgages, and corporations start canceling software subscriptions en masse due to collapsing demand, the house of cards collapses. Algorithms began synchronously dumping the securities of any companies whose business model relied on stable middle-class wages, turning the correction into an unmanageable peak.

The crisis is becoming global. India is on the verge of default, its IT sector (TCS, Infosys, Wipro) loses export revenue of $200 billion. Indian programmers cannot compete on price: the marginal cost of an AI coding agent has fallen to the cost of electricity. Rupee collapses 18%, IMF starts emergency talks with Delhi.

June 2028: Current. Unemployment jumped to 10.2%, the S&P 500 collapsed 38% from its peak. The real estate bubble burst in San Francisco, Seattle and Austin - prices fell 8-11%. White collar incomes were the foundation of the $13 trillion mortgage market, and now that foundation has cracked.

While authorities discuss imposing a tax on computing power and paying an "AI dividend" to support millions of laid-off workers, Occupy Silicon Valley activists are blocking the offices of Anthropic and OpenAI.

All the profits from the productivity boom go to the owners of powerful servers and the shareholders of laboratories, pushing social inequality to the limit.

"But it's not June 2028 on the calendar. This text is written in February 2026. The S&P 500 index is holding at historic peaks. Negative feedback loops have not yet begun to tighten. We believe that some of these scenarios will never materialize. However, we are equally confident: the potential for machine intelligence will continue to grow exponentially. The market markup that human intelligence receives today will inevitably shrink"

Citrini presents his scenario as a warning: investors should not succumb to a false sense of security when looking at the abundance of AI deals. Right now, the market is cheerfully welcoming corporate downsizing because it inflates corporate margins, but it ignores the base - the economy lives off the spending of those very people.

Once white-collar incomes can no longer keep pace with rising prices, no amount of AI hype can hide the collapse of consumer demand. To avoid a systemic collapse, the authors call for preparing for a "participatory economy" model, where taxes on computing and direct dividends to citizens will become a necessity. Investors are advised to hedge their risks now, before corporate profits become completely disconnected from the reality of empty pockets of consumers.

This article was AI-translated and verified by a human editor