The AI Boom's Hidden Risk: A Repeat of Japan's 'Lost Decades'

A troubling pattern in economic history is repeating itself, and artificial intelligence is the latest catalyst. Throughout modern history, from the power looms of 19th-century Manchester to the automation of 1970s Detroit, transformative technologies have followed a predictable script. Initial productivity surges and market euphoria give way to a long, quiet stall as wage stagnation weakens overall demand. We may be in the opening act of that play once more.

Japan’s experience from 1990 onward is the clearest blueprint. After its asset bubble burst, the country entered decades of stagnation despite high productivity and corporate profitability. The core failure was a collapse in demand; households and businesses saved instead of spent. The Bank of Japan’s extraordinary monetary measures proved powerless to stimulate an economy where incomes weren't rising. The Nikkei stock index took 35 years to recover its 1989 peak. Tokyo real estate values fell 70% and never fully came back.

This ‘productivity trap’ is now a risk for AI-driven economies. The critical difference this time is the target: cognitive workers. Programmers, analysts, and managers—the professional class with high spending power—are facing displacement. Previous technological shocks affected manual labor; AI can replicate knowledge work globally at near-zero cost. The new jobs being created often pay less than the roles being erased, creating a severe threat to aggregate demand.

For investors, this suggests a potential regime shift. The coming years may not bring a dramatic crash, but a grinding, volatile stagnation similar to Japan’s experience or the U.S. markets of 1968-1982. In such an environment, passive index investing could fail. Extreme dispersion between a few winning stocks and many laggards would make stock-picking essential. Cash would transform from a drag to a strategic tool for seizing frequent dislocations, while assets like gold could hedge against policy uncertainty.

The narrative of limitless tech growth may fracture slowly. If AI’s primary corporate use is cost-cutting, the very customers for AI services may shrink their own spending. Markets are currently pricing in the immediate productivity gains but ignoring the second-order effect of weakened demand. The lesson from history is that productivity alone isn't enough. Without a mechanism to circulate income and sustain demand, even the most advanced technology can lead to an era of sideways markets and eroded returns. The policy response to this will be critical, but history shows it often arrives painfully late.