The 4 Things You Need for a Tech Bubble
The tech world is abuzz with speculation, a collective murmur growing into a loud debate: Is artificial intelligence fueling an economic bubble on par with, or even exceeding, the infamous dot-com crash? This pressing question formed the core of a recent Uncanny Valley episode from WIRED, where hosts Michael Calore and Lauren Goode were joined by WIRED contributor Brian Merchant, author of the newsletter Blood in the Machine. Merchant delved into a historical framework designed to identify the classic signs of an economic bubble, applying it directly to the current AI craze and exploring its profound implications for everyone.
The conversation highlighted the staggering scale of current AI investments. Tech giants like Google, Meta, and Microsoft are not just participating; they are doubling down, pledging massive capital expenditures (CapEx) for AI infrastructure, with commitments extending to 2026. This year alone, big tech is on track to pour an estimated $400 billion into artificial intelligence, and even that is deemed insufficient by some. The market has seen unprecedented valuations, with chipmaker Nvidia becoming the world’s first $5 trillion company, a figure 2.5 times the entire Canadian economy. Yet, amidst this frenetic investment, a stark reality emerges: reports indicate that 95 percent of businesses currently utilizing AI claim to be seeing little to no return on their investment. This stark discrepancy begs the question: are we merely witnessing a natural technological evolution, or are we teetering on the brink of an unprecedented economic bubble?
To dissect this complex issue, Merchant turned to the analytical framework developed by scholars Brent Goldfarb and David Kirsch, outlined in their 2019 book, Bubbles and Crashes: The Boom and Bust of Technological Innovation. After meticulously analyzing numerous historical tech bubbles, Goldfarb and Kirsch identified four critical factors that consistently contribute to the formation and eventual burst of these economic phenomena. Applying this robust framework, Merchant posited that AI not only exhibits these characteristics but does so with an intensity that suggests a "bubble to end all bubbles."
The first crucial factor is Uncertainty in Innovation. A tech bubble often forms when there is a groundbreaking new technology or innovation whose long-term business model, revenue generation, and market applications are fundamentally unclear. While the technology itself may be awe-inspiring and undeniably powerful, investors struggle to define precisely how it will translate into sustained profitability. Merchant illustrated this with historical examples such as the early days of electricity and radio. When electric light first emerged in the 19th century, its transformative potential was evident. Streets could be illuminated, and powerful demonstrations captivated the public. However, the path to commercial viability was fraught with uncertainty. Would the primary market be home light bulbs or large municipal lighting fixtures? What would be truly cost-effective? It took decades for these questions to be resolved and for the technology to transition from laboratory marvel to widespread commercial success. Similarly, early radio broadcasting, despite its revolutionary nature, lacked clear vectors for a business model, leading to an investment boom in the 1920s that ultimately became a bubble.
Today, AI mirrors this uncertainty. While chatbots like ChatGPT have gone viral, demonstrating immense consumer interest and technical prowess, the fundamental question of how AI will consistently generate substantial, sustainable profits remains largely unanswered. Companies are pouring billions into AI infrastructure, yet, as the MIT study highlighted, the vast majority of businesses deploying generative AI as an automation tool have yet to see tangible returns. Sam Altman, CEO of OpenAI, famously quipped years ago that their business plan was to "build AGI and ask it" how to make money. This anecdote, while humorous, underscores the profound ambiguity surrounding AI’s commercial future, making it a prime candidate for a bubble fueled by uncertainty. The promise of "limitless potential" often serves to obscure the absence of clear, viable business models.
The second factor is the presence of Pure Play Investments. These are companies whose fate is inextricably tied to the business success of a single innovation or technology. Their entire existence and valuation hinge on that particular technology fulfilling its hyped potential. In the current AI landscape, Nvidia stands out as the quintessential pure-play. Prior to the AI boom, Nvidia was primarily known for its graphical processing units (GPUs) used in gaming. However, it has strategically pivoted and foregrounded its entire business into supplying the essential chips that power the AI revolution. It’s the classic "selling shovels during a gold rush" scenario, making it appear as the most certain bet in an otherwise uncertain market. Companies like CoreWeave, which specialize in renting out cloud compute space for AI, are further examples of pure-play investments whose survival is directly linked to the sustained growth of the AI market. The sheer concentration of capital in such companies, with Nvidia alone constituting approximately 8 percent of the entire stock market, amplifies the risk if the AI bubble were to burst.
The third critical element for a tech bubble is the involvement of Novice Investors. This refers to retail investors or non-experts who gain easy access to investment vehicles tied to the new innovation, often without fully understanding the underlying fundamentals or risks. The dot-com boom famously saw ordinary individuals pouring money into nascent internet companies based on general enthusiasm rather than deep market analysis. Today, platforms like Robinhood make it incredibly simple for anyone to invest in publicly traded companies like Nvidia. However, Merchant and Goldfarb argue that in the context of AI, the concept of "novice investor" expands significantly. Due to the unprecedented level of uncertainty surrounding AI’s future applications and profitability, even seasoned institutional investors find themselves operating with a degree of novice-like speculation.
Furthermore, the public’s exposure to an AI bust is not limited to direct stock purchases. There’s a growing web of "circular investing" that indirectly ties diversified portfolios and pension funds to the AI boom. For instance, real estate companies investing heavily in data centers become indirectly dependent on the AI sector’s growth. High-profile deals, such as OpenAI’s stake in chipmaker AMD, mean that a bust in OpenAI could have ripple effects on AMD’s valuation, impacting its public shareholders. This pervasive, often indirect, exposure to AI-related investments means that a potential collapse could affect a much broader segment of the population, far beyond those actively trading AI stocks. The sheer scale of institutional money involved, combined with the inherent unknowability of AI’s long-term trajectory, elevates the risk for virtually everyone with investments.
Finally, the fourth factor is a Coordinating Belief or an Alignment of Beliefs among investors that a particular innovation is destined to be "the future." This shared narrative is often solidified by real-world use cases or demonstrations that capture public imagination and reinforce confidence. Charles Lindbergh’s transatlantic flight, for example, served as a monumental tech demo for the nascent aviation industry, galvanizing investor belief and leading to an aviation bubble. In the current era, ChatGPT’s viral launch provided a similar, organic demonstration of AI’s capabilities, fostering a widespread conviction that "this is the real deal."
The narrative surrounding AI is particularly potent, promising solutions to nearly every conceivable challenge. It’s touted as a tool to automate jobs, cure cancer, fight climate change, and even usher in an era of Artificial General Intelligence (AGI) capable of performing any human task. This promise of "limitless potential" is incredibly convenient, as it allows for a broad and often vague definition of success, making it appealing to diverse investor interests. Whether an investor is looking for labor automation, pharmaceutical breakthroughs, or environmental solutions, AI’s expansive narrative offers something for everyone. This powerful, coordinated belief system, reinforced by impressive (though often limited) demos and the fear of missing out (FOMO), has driven billions into the sector, despite the underlying uncertainties.
When Goldfarb and Kirsch applied their 0-8 scale to AI, they returned a "big fat eight," indicating the maximum level of bubble alert. This assessment suggests that all four ingredients for a tech bubble are not only present but are manifesting in significantly high doses. While their framework primarily identifies the existence of a bubble, Merchant went further, cautioning that the scale of this potential AI bubble could lead to economic calamity far surpassing previous busts, including the dot-com crash.
Despite the dire warnings, not everyone agrees. Nvidia CEO Jensen Huang, for instance, dismisses the bubble talk, characterizing the current landscape as a "natural transition from general purpose computing to accelerated computing." This perspective highlights the inherent difficulty in identifying a bubble in real-time, as the answer often depends on who you ask and their vested interests.
Even if the AI bubble were to burst, history suggests that technological innovations rarely vanish entirely. The internet continued its exponential growth after the dot-com crash, eventually becoming an indispensable part of modern life. Similarly, AI will undoubtedly leave an enduring legacy. The utility found post-burst, however, might not align with the utopian promises. Merchant suggested that AI’s primary long-term utility might lie in content production and labor automation – a "wage reduction tool." This could lead to a future where AI-generated content fills our digital feeds, and AI avatars handle customer service or even managerial roles. Such an outcome, while demonstrating persistent utility, raises concerns about its long-term impact on the workforce and society.
Another unique aspect of the current situation is the potential for unprecedented political intervention. With governments recognizing the strategic importance of AI, and even taking stakes in technology companies (as the US administration has done with Intel), a future scenario could see states economically intervening to prop up AI firms should they face a financial crisis. This adds another layer of complexity to predicting the fallout, suggesting that the traditional boom-and-bust cycle might be mitigated or reshaped by political will.
Ultimately, the question of whether AI is a bubble, and what its eventual fate will be, remains one of profound uncertainty. The massive investments, the unclear business models, the widespread investor exposure, and the compelling narrative all point to an unprecedented economic phenomenon. While some utility will undoubtedly persist, the scale of potential economic pain and the long-term societal ramifications of AI’s integration into our lives make it a topic that demands vigilant observation and critical analysis.
