There’s an ongoing debate about whether artificial intelligence is in a bubble. The more important question is whether that’s necessarily a bad thing.
The seemingly irrational bubbles of capitalism have historically been more effective at driving innovation than the carefully planned strategies of socialist systems. Without massive waste, there are no technological breakthroughs. That is precisely the fundamental strength of market economies compared to centrally planned ones.
Innovative bubbles
It’s important to distinguish between financial and innovative bubbles. Financial bubbles, like those leading up to the Great Financial Crisis, mainly involved inflating existing assets without creating new value. In contrast, innovative bubbles, such as the dot-com bubble or the 19th-century railway mania, produced real infrastructure and technological breakthroughs.
Still, 99 percent of dot-com companies no longer exist. Pets.com, Webvan, and eToys.com burned through billions in capital. Yet the infrastructure built during that frenzy—fiber-optic cables, server farms, internet protocols, and payment systems—became the foundation of today’s digital economy. Without those “insane” overinvestments, we wouldn’t have smartphones streaming video, cloud computing, or e-commerce. Telecom companies laid so much fiber that data prices fell by more than 99 percent. What was a disaster for individual investors became a blessing for society.
Mobilizing illusion
The fact that successful innovations are quickly copied should discourage innovation. But bubbles overcome this restraint by creating the perception of enormous potential profits. That illusion mobilizes capital on a scale that rational analysis would never justify.
In socialist systems, a committee must decide in advance which technologies seem promising. But no central authority has enough information to make such complex decisions. The Soviet Union demonstrated this painfully. Despite launching the first satellite, it fell hopelessly behind in consumer electronics, computers, and the internet. The central planning bureau, Gosplan, allocated resources based on political priorities rather than market potential.
China’s technological rise only began when it allowed private entrepreneurship. Alibaba, Tencent, and ByteDance did not emerge from government planning but from quasi-capitalist dynamics. When Xi Jinping began to tighten control, innovation slowed. The state-led push to dominate semiconductors has so far failed to produce a breakthrough comparable to Taiwan’s TSMC, despite hundreds of billions in investments.
Distributed knowledge
The “wisdom of crowds” during a bubble mobilizes the distributed knowledge of thousands of investors and entrepreneurs, each seeing a piece of the puzzle. During the dot-com bubble, thousands of companies experimented in parallel with different business models. Most failed, but those experiments produced today’s winning formulas: ad-supported search engines, online marketplaces with reviews, social networks, and video streaming.
No central planner could have tried all these approaches at once. That said, China now has more than a hundred electric vehicle manufacturers, only a small fraction of which are profitable. It seems that China, too, is drawing on the successful formulas of capitalism.
Britain’s 19th-century railway bubble illustrates the success of market-driven rather than centrally planned innovation. The outcome was chaotic—parallel lines and even experiments with rocket-powered locomotives. France’s railway system, elegantly planned under Napoleon III, was more efficient on paper. But Britain’s chaos-driven system spurred more innovation and produced signaling, rail, and locomotive designs adopted worldwide.
The Soviet Union had brilliant scientists, but without market incentives, innovations stayed in the lab. In the 1960s, Viktor Glushkov developed Ogas, a national computer network similar to Arpanet. But while Arpanet evolved into the internet, Ogas died quietly in bureaucracy. The Soviets even developed the first mobile phone in 1957, but without commercial incentives, it remained a curiosity.
The current AI bubble
The current AI bubble shows the same dynamic. Tech giants are investing tens of billions in data centers that may never be fully used. Hundreds of startups are developing competing AI models, burning through millions in computing costs. From a planning perspective, this looks like madness. But all those “excess” data centers will be available for applications we can’t yet imagine—just as the dot-com fiber network eventually enabled Netflix.
The superiority of the capitalist innovation model lies in its capacity to be temporarily irrational. Central planning cannot create a bubble because no bureaucrat can get rich by believing in a five-year plan. But the “animal spirits” of greed and Fomo can mobilize the waves of investment required for breakthroughs.
Capitalism accepts waste as the price of innovation. Funding hundreds of projects, 99 percent of which fail, may seem absurd—but that “inefficiency” is precisely its strength. Socialism may distribute what already exists more fairly, but capitalism is superior at creating what doesn’t yet exist. Waste is not an unfortunate byproduct but the necessary price of progress—a price that, considering the historical alternatives, is well worth paying.
Han Dieperink is chief investment officer at Auréus Vermogensbeheer. He previously served as chief investment officer at Rabobank and Schretlen & Co.