AI's High-Stakes Bubble: What History Teaches Us

Summary

The current AI boom is the largest financial bet in history, with $700 billion invested this year and projected to reach $6.7 trillion by 2030. Despite massive spending, most users access AI for free, and companies are accumulating significant debt while generating minimal revenue, yet valuations continue to soar. History reveals a recurring pattern with revolutionary technologies: an initial wave of investors funds massive infrastructure buildouts, only to face financial ruin as the long-term revenue doesn't match the upfront costs and debt. Examples include the UK canals, British and American railways, and fiber optic cable. These early investors are often wiped out, but the infrastructure they financed ultimately powers economic growth. AI presents a unique challenge because its most expensive components, like GPUs, have a short lifespan of about 3 years, unlike the durable infrastructure of past revolutions. This rapid obsolescence creates a permanent 'tax' on innovation, unlike previous technologies where the core infrastructure lasted decades. Furthermore, concerns are rising that AI companies might be inflating chip depreciation schedules to hide significant losses, potentially around $176 billion. This practice, similar to financial engineering used in the 2008 crisis, could lead to banks offloading AI data center debt to pension funds and insurers, while IPOs allow insiders to exit at inflated valuations. The recommended investment strategy is to bet on the AI sector rather than individual companies, maintain a long-term perspective, avoid debt-financed speculation, and diversify across economic forces due to the inherent risks and the system's tendency to disadvantage novice investors.