According to author Bernd Hobart, bubbles aren't just irrational speculation. Sky-high valuations signal to all players—from power plants to chip fabs to software developers—that the "time is now." This encourages massive, parallel investments that might otherwise be too risky, effectively manufacturing the future just in time.
Related Insights
Speculative manias, like the AI boom, function like collective hallucinations. The overwhelming belief in future demand becomes self-fulfilling, attracting capital that builds tangible infrastructure (e.g., data centers, fiber optic cables) long before cash flows appear, often leaving lasting value even after the bubble bursts.
The memo argues that the "hysteria of the bubble" compresses the timeline for building out new technologies from decades into just a few years. Patient, value-focused investing would never fund the massive, parallel, and often wasteful experimentation required to jump-start a new technological paradigm at such a rapid pace.
The current AI investment surge is a dangerous "resource grab" phase, not a typical bubble. Companies are desperately securing scarce resources—power, chips, and top scientists—driven by existential fear of being left behind. This isn't a normal CapEx cycle; the spending is almost guaranteed until a dead-end is proven.
The current AI spending frenzy uniquely merges elements from all major historical bubbles—real estate (data centers), technology, loose credit, and a government backstop—making a soft landing improbable. This convergence of risk factors is unprecedented.
The current AI investment frenzy is a powerful feedback loop. Silicon Valley labs promote a grand narrative to justify huge capital needs. Simultaneously, Wall Street firms earn massive fees by financing this buildout, creating a shared, bi-coastal incentive to keep the 'super cycle' narrative going, independent of immediate profitability.
The most immediate systemic risk from AI may not be mass unemployment but an unsustainable financial market bubble. Sky-high valuations of AI-related companies pose a more significant short-term threat to economic stability than the still-developing impact of AI on the job market.
A genuine technological wave, like AI, creates rapid wealth, which inherently attracts speculators. Therefore, bubble-like behavior is a predictable side effect of a real revolution, not proof that the underlying technology is fake. The two phenomena come together as a pair.
Current AI spending appears bubble-like, but it's not propping up unprofitable operations. Inference is already profitable. The immense cash burn is a deliberate, forward-looking investment in developing future, more powerful models, not a sign of a failing business model. This re-frames the financial risk.
Howard Marks distinguishes between two bubble types. "Mean reversion" bubbles (e.g., subprime mortgages) create no lasting value. In contrast, "inflection bubbles" (e.g., railroads, internet, AI) fund the necessary, often money-losing, infrastructure that accelerates technological progress for society, even as they destroy investor wealth.
Leaders from NVIDIA, OpenAI, and Microsoft are mutually dependent as customers, suppliers, and investors. This creates a powerful, self-reinforcing growth loop that props up the entire AI sector, making it look like a "white elephant gift-giving party" where everyone is invested in each other's success.
