🔗 Share this article

The California Gold Rush forever altered the American story. Between 1848 to 1855, some 300,000 people descended there, lured by dreams of riches. This migration came at a devastating cost, involving the displacement of Indigenous peoples. Yet, the true winners were often not the prospectors, but the businessmen selling them shovels and canvas trousers.

Today, California is witnessing a new kind of rush. Centered in Silicon Valley, the new prize is AI. This central debate is no longer whether this is a speculative bubble—numerous voices, from AI insiders and central banks, argue it clearly is. Instead, the real inquiry is understanding what kind of bubble it is and, crucially, the enduring impact will be.

The History of Manias and Their Aftermath

All speculative frenzies exhibit a key characteristic: speculators pursuing a vision. But their manifestations vary. During the late 2000s, the housing bubble almost collapsed the world financial system. Before that, the dot-com bubble burst when investors realized that web-based grocery retailers were not inherently profitable.

This pattern extends far back. From the 17th-century Netherlands tulip mania to the 18th-century South Sea bubble, history is replete with cases of irrational exuberance ending in collapse. Research suggests that virtually every major investment frontier invites a investment surge that ultimately goes too far.

Almost every emerging domain made available to investment has resulted in a financial bubble. Investors have scrambled to tap into its potential only to overshoot and stampede in panic.

The Critical Distinction: Housing or Housing?

Thus, the essential issue about the AI funding frenzy is not about its eventual deflation, but the nature of its aftermath. Would it mirror the housing bubble, which left a crippled banking sector and a deep, protracted downturn? Or, could it be more like the dot-com crash, which, although disruptive, in the end paved the way for the modern digital economy?

A major determinant is financing. The housing bubble was fueled by reckless mortgage credit. The current concern is that this AI-driven spending spree is also reliant on debt. Major tech firms have reportedly raised unprecedented sums of debt this period to fund expensive data centers and chips.

Such reliance creates broader vulnerability. If the bubble deflates, heavily indebted entities could fail, possibly causing a credit crunch that extends well past Silicon Valley.

An A Deeper Question: Is the Technology Even Viable?

Apart from finance, a even more basic uncertainty exists: Can the prevailing approach to AI itself endure? Past bubbles frequently left behind transformative infrastructure, like railways or the web.

However, prominent thinkers in the field now doubt the path. Experts argue that the massive investment in LLMs may be misplaced. They contend that achieving genuine AGI—the superhuman intelligence—demands a radically different approach, like a "world model" design, instead of the existing statistical models.

If this perspective turns out to be correct, a sizable chunk of today's astronomical technology investment could be channeled toward a technological dead end. Similar to the 49ers of old, today's backers might discover that selling the shovels—here, chips and computing capacity—does not guarantee that you'll find real gold to be unearthed.

Final Thought

The artificial intelligence chapter is certainly a speculative surge. The critical work for observers, policymakers, and the public is to see past the inevitable market correction and focus on the two legacies it will create: the financial wreckage of its aftermath and the practical assets, if any, that remain. Our long-term may well depend on which legacy ends up more significant.