No one is buying a Sony Discman in 2026, and no one is going to argue that it would be a good idea. Why? Because “creative destruction” is critical to innovation. Capitalism is the most productive system for creating value because it allows people to invest their wealth toward the opportunities where they sense the highest return, which are often at the forefront of innovation. In the process, the most valuable industries and technologies receive funding, displacing less productive sectors. Instead of MP3 players, many people are investing in new technologies.
Right now, investors are betting on a massive AI-driven boom in economic productivity. Many expect the efficiency gains to be so large that they could actually bring about deflation. While “AI” and large language models like ChatGPT are often used synonymously, the potential benefits of the technology go far beyond chatbots, from curing diseases to automated weed detection and elimination. Trillions of dollars are now being allocated toward AI, and neither individual investors nor the tech companies themselves would commit capital on this scale without expecting a huge return on investment, creating immense value for society in the process.
That being said, the same creative destruction that enables innovation can also lead to the desolation of industries that don’t provide real value.
In 1840s Britain, everyone knew that railroads were going to revolutionize the world. As such, immense amounts of capital poured into the markets to finance new lines. Parliament considered hundreds of railway bills in 1846 alone and authorized thousands of miles of track between 1845 and 1847. Then credit tightened. The Bank of England raised discount rates in October 1845, and over the next five years, the collective value of publicly traded railway securities fell 64 percent.
The same cycle repeated itself in the United States one generation later. The Panic of 1873 was triggered by the collapse of Jay Cooke & Co., an investment bank that had purchased large quantities of railroad bonds. In the wake of the panic, it became clear that many rail projects were redundant lines to nowhere, now referred to as “ghost tracks.” The banker James Lees had seen it coming, warning nine months earlier of
capital “wasted in extravagance and ill spent in wildcat enterprises such as railroads through deserts, beginning nowhere and ending nowhere.”
Trains changed the world, but not in the way that people expected. The railroad companies were not the primary beneficiaries. Instead, the primary beneficiaries were the industries that benefited from the productivity boom enabled by rail travel. In the end, the useful tracks remained while investors absorbed losses on the rest.
The relevant lesson from both of these manias is that there’s little reason for Americans to worry about data centers broadly, especially when the current data shows that fears about AI-induced unemployment are overblown. As long as this technology is not subsidized by communities and taxpayers, new digital infrastructure will either create immense value for society, or it will simply be an overbuild, just like the train tracks to nowhere.
If companies like OpenAI fail to generate real profits and data centers turn out to be a massive waste of money, investors will wreak losses and capital will be reallocated elsewhere. If the opposite is true, and we need an unimaginable amount of processing power to enable the Fourth Industrial Revolution, society will benefit from a surplus of data centers.
The broader point is that we need to let this process play out in order to understand where the true value of AI lies. No one can predict the future, and capital markets should be allowed to engage in value discovery without interruption in order to help humanity progress.
