Historical data shows that when CapEx for a new technology exceeds 2-3% of GDP, a market crash follows within a few years. Today's AI infrastructure spending has reached similar levels, with 93% of GDP growth coming from AI CapEx, suggesting the current tech boom is unsustainable and headed for a correction.

The AI bubble resembles the telecom bubble of the late 90s, where massive, real CapEx on physical infrastructure (fiber optic cables then, GPUs now) created real profits for suppliers. The danger is this euphoria, funded by cheap capital, leads to overinvestment with no guarantee of long-term profitability.

The current AI-driven CapEx cycle is analogous to historical bubbles like the 19th-century railroad buildout and the dot-com boom. These periods of intense capital investment have historically led to major economic downturns and secular bear markets, suggesting a grim multi-year outlook beyond the current cycle.

Massive upfront capital expenditure (CapEx) for AI infrastructure creates a timing gap before revenue materializes. This mirrors historical bubbles like the dot-com and railroad eras, where the technology succeeded but early investors were wiped out waiting for returns.

Tech giants are spending hundreds of billions on AI infrastructure with slow initial results, reminiscent of the Web 1.0 era's overbuild of fiber optic networks. This parallel suggests a potential AI bubble where the infrastructure is built, but the equity holders who funded it get crushed in a market correction.

The tangible economic effect of the AI boom is currently concentrated in physical capital investment, such as data centers and software, rather than widespread changes in labor productivity or employment. A potential market correction would thus directly threaten this investment-led growth.

The massive capital expenditure in AI infrastructure is analogous to the fiber optic cable buildout during the dot-com bubble. While eventually beneficial to the economy, it may create about a decade of excess, dormant infrastructure before traffic and use cases catch up, posing a risk to equity valuations.

The massive spending on AI infrastructure may be a form of 'malinvestment,' similar to the telecom buildout during the dot-com boom. Rajan warns that while AI's promise is real, the transition from infrastructure creation to widespread, profitable use could be slow, creating a valuation gap and risk of a market correction.

Unlike past tech bubbles built on unproven ideas, AI technology demonstrably works. The systemic risk lies in the unprecedented capital expenditure by hyperscalers on data centers, reminiscent of the "dark fiber" overinvestment during the telecom bubble. A demand shortfall for this new capacity is the real threat to the economy.