Jim Chanos: The AI Bubble Is “Much Worse” Than Dot-Com

The current AI infrastructure boom mirrors the late 1990s dot-com era, characterized by speculative fervor where market participants prioritize promises over economic reality. Massive capital expenditures on data centers and AI hardware are frequently financed through debt and equity, despite uncertain long-term returns and questionable profitability. While NVIDIA serves as a critical gatekeeper, many dependent companies trade at unjustifiably high valuations, reflecting significant market dispersion. Furthermore, aggressive accounting practices, such as classifying assets as "construction in progress," allow firms to defer depreciation, obscuring the true financial burden of these projects. As interest rates remain elevated, the mismatch between long-term capital commitments and near-term spot pricing creates systemic risk, particularly as the return on incremental invested capital for major hyperscalers continues to decline.