While many early investors in tech booms (e.g., telecom, AI) lose money, these 'bubbles' are not a societal waste. They fund the rapid construction of foundational infrastructure, like fiber optic networks or data centers, creating immense long-term value and options for future innovation that society ultimately benefits from.
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.
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.
Bubbles provide the capital for foundational technological shifts. Inflated valuations allow companies like OpenAI to raise and spend astronomical sums on R&D for things like model training, creating advances that wouldn't happen otherwise. The key for investors is to survive the crash and back the durable winners that emerge.
History shows pioneers who fund massive infrastructure shifts, like railroads or the early internet, frequently lose their investment. The real profits are captured later by companies that build services on top of the now-established, de-risked platform.
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 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 epicenter of a tech boom is rarely the new technology itself. Instead, capital floods into adjacent, understandable sectors. The dot-com bubble wasn't about software but a massive telecom infrastructure bubble, fueled by debt financing for tangible assets like fiber and buildings.
The current AI infrastructure build-out avoids the dot-com bubble's waste. In 2000, 97% of telecom fiber was unused ('dark'). Today, all GPUs are actively utilized, and the largest investors (big tech) are seeing positive returns on their capital, indicating real demand and value creation.
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.
The massive capex spending on AI data centers is less about clear ROI and more about propping up the economy. Similar to how China built empty cities to fuel its GDP, tech giants are building vast digital infrastructure. This creates a bubble that keeps economic indicators positive and aligns incentives, even if the underlying business case is unproven.
