Citrini Research's widely shared report, "The 2028 Global Intelligence Crisis," has generated a lot of discussion about the economy's future. The study examines a hypothetical apocalyptic scenario in which widespread unemployment and financial instability are caused by fast AI automation by 2028. It explicitly cautions that AI-driven automation could destroy the business models of Indian IT behemoths like TCS, Infosys, and Wipro.
According to the paper, clients began employing AI coding agents at a fraction of the cost, which caused India's IT services industry, which exported more than $200 billion a year, to collapse by 2028.One value proposition served as the foundation for the entire model: Indian coders are far less expensive than their American counterparts.

However, "the marginal cost of an AI coding agent had collapsed to, essentially, the cost of electricity," the paper said. "The rupee falls significantly against the dollar within four months as services exports weaken."Contract cancellations at TCS, Infosys, and Wipro increased until 2027. The services surplus that had supported India's external accounts vanished, causing the rupee to drop 18% versus the dollar in just four months. The IMF had started "preliminary discussions" with New Delhi by the first quarter of 2028.

Absence of a Natural Brake
According to the report, even if consumers were spending less money, businesses continued to invest in AI because it created a feedback loop without a natural brake.AI improved and became more affordable. After laying off employees, businesses used the money saved to purchase more AI capabilities, which allowed them to fire even more employees. It stated, "Workers who were displaced spent less."To preserve margins, consumer-facing businesses sold fewer products, weakened, and increased their investments in AI. AI improved and became more affordable. a feedback cycle without a built-in brake.
The irony of the scenario was that major AI companies like NVIDIA and TSM continued to perform as the economy they were upending started to decline.

"NVDA continued to report record profits. TSM continued to operate at a utilization rate of 95% or above. The hyperscalers continued to spend $150–200 billion on data center capital expenditures each quarter. Taiwan and Korea, two economies that were solely convex to this trend, performed far better.