Tech NewsThoughtsWe're living in the storytelling economy. The numbers are staggering, almost incomprehensible: tech companies are projected to spend about $400 billion this year on AI infrastructure, more than the inflation-adjusted cost of putting humans on the moon. Yet consumers spend a mere $12 billion annually on AI services. If you understand the economic chasm between Singapore's GDP and Somalia's, you grasp the disconnect between AI's vision and reality.
The question isn't whether we're in an AI bubble—even industry leaders admit we are. The real questions are: How catastrophic will the burst be? And perhaps more surprisingly, could it actually help us?
AI-related capital expenditures surpassed U.S. consumer spending as the primary driver of economic growth in the first half of 2025, accounting for 1.1% of GDP growth. This isn't just impressive—it's unprecedented and unsustainable. JP Morgan's analysis shows that AI-related stocks have accounted for 75% of S&P 500 returns, 80% of earnings growth, and 90% of capital spending growth since ChatGPT launched in November 2022.
The concentration of wealth and risk is staggering. The weighted-average share price of the Magnificent Seven corporations increased by 156% during 2023-2024, while the other 493 firms in the S&P 500 experienced an average increase of just 25%. These seven companies now control more than one-third of the entire index—a concentration that violates every diversification principle we learned from the 1929 crash.
The deals themselves have become Byzantine. OpenAI is taking a 10% stake in AMD while Nvidia invests $100 billion in OpenAI, which also counts Microsoft as a major shareholder, even as Microsoft remains a major customer of CoreWeave, where Nvidia holds significant equity. In less than three years, OpenAI has transformed from a research project into a pillar of the global economy, with Oracle committing to a $300 billion, five-year computing deal—despite OpenAI losing billions of dollars annually while projecting only $13 billion in revenues for 2025.
While the hype machine churns, reality tells a different story. An August 2025 MIT report concluded that 95% of generative AI pilot projects in companies are failing to raise revenue growth. The promised productivity revolution hasn't materialized. A Danish study of 25,000 workers across 7,000 workplaces found that AI chatbots have had no significant impact on earnings or recorded hours in any occupation, with modest productivity gains averaging just 3% time savings.
Companies that rushed to replace workers with AI are now backpedaling. Klarna, which boasted in March 2024 that its AI assistant was doing the work of 700 laid-off workers, rehired them as gig workers by summer 2025. IBM followed a similar pattern, reemploying staff after laying off about 8,000 workers for automation. The vision of AI revolutionizing productivity is colliding with the hard truth that generic tools like ChatGPT stall in enterprise use since they don't learn from or adapt to workflows.
Comparisons to the dot-com bubble are inevitable but imperfect. The International Monetary Fund's chief economist warns that the AI investment boom has echoes of the 1990s dot-com bubble, though the current scale is smaller, with AI-related investment increasing by less than 0.4% of U.S. GDP since 2022, compared with the dot-com era's 1.2% increase between 1995 and 2000.
But there's a crucial difference that makes this potentially more dangerous: opacity. Big AI firms are shifting huge amounts of spending off their books into special purpose vehicles (SPVs) that disguise the cost of the AI build-out. The accounting tricks used to depress reported infrastructure spending while inflating profits echo the financial engineering that preceded the 2008 housing crisis.
Yet there are also protective factors. Unlike the housing bubble, the AI boom is mainly being financed by companies with sizable cash holdings and strong balance sheets rather than leverage, which could limit contagion when valuations correct.
Perhaps the most insidious effect of the AI bubble isn't what it's building, but what it's starving. Financial expert Paul Kedrosky draws a chilling parallel: During the 1990s, massive telecom capital spending diverted investment away from U.S. manufacturing, starving small manufacturers of capital and increasing their cost of capital, inadvertently contributing to the loss of manufacturing jobs as they struggled to compete with China.
History is repeating itself. If you're a small manufacturer hoping to benefit from on-shoring due to tariffs, the hurdle rate for raising capital has become much higher because investors are comparing you to AI's seemingly tremendous returns. The AI "death star" is sucking capital out of every other sector of the economy, from healthcare to infrastructure to clean energy.
What happens when this bubble bursts? The immediate effects will be severe. A sharp contraction in tech would put tens of thousands of people out of work, vaporize trillions of investment dollars, torpedo retirement and education funds, obliterate life savings, and ruin lives. Dormant data centers could become the new abandoned shopping malls, monuments to speculative excess.
Anthropic CEO Dario Amodei warned that AI could wipe out half of all entry-level white-collar jobs and spike unemployment to 10-20% in the next one to five years. Yet this prediction exists in tension with the current reality: AI isn't replacing workers as promised, and when it does, companies are often rehiring them.
The economic indicators are already flashing warning signs. U.S. youth unemployment has risen from 6.6% to 10.5% since April 2023, with job vacancies for career starters falling by more than 30%. Meanwhile, recent analysis from Moody's shows that all real spending growth over the last year has come from the top quintile of the income distribution, with everyone else just treading water.
Here's where the story takes an unexpected turn. Some economists argue that the bubble bursting might actually benefit most Americans. Dean Baker of the Center for Economic and Policy Research uses a bathtub metaphor: The "rich people" faucet (AI spending) is gushing while the "ordinary workers" faucet barely trickles; when the bubble bursts and the water level drops, it creates room to turn the flow higher from the ordinary workers faucet without causing inflation.
In other words, the collapse would free up economic space for policies that benefit regular people—if policymakers seize the opportunity. The Federal Reserve could lower interest rates, and the government could increase spending on healthcare, education, childcare, and green transition subsidies. The bubble's collapse would create the room the economy needs for such interventions.
This isn't guaranteed, of course. After both the dot-com and housing bubbles burst, political obstacles prevented adequate government response, leading to prolonged periods of high unemployment. After the tech bubble burst in 2000-01, it took four full years to regain lost jobs—the longest period without job growth since the Great Depression.
No two bubbles are exactly alike, and the most important question is what remains after they pop. The dot-com bubble left behind genuine infrastructure and innovations that enabled the modern internet economy. The tulip mania left only cautionary tales and financial ruin.
Where will AI fall on this spectrum? The technology does work—for certain tasks. Developers use it to debug code, writers use it to transcribe audio, and researchers use it to summarize papers. But these practical applications are far cry from the transformative vision being sold to justify trillion-dollar valuations.
While a lot of fanciful dot-com ideas eventually went away, we're more online now than ever; the same kind of distillation might play out for AI. Useful applications will survive and improve, while the hype-driven nonsense will vanish. As technology writer Cory Doctorow notes, after the burst we can expect cheap GPUs, a buyer's market for data scientists, and promising open-source models with vast potential for improvement.
Ultimately, public sentiment may determine whether the current AI boom proves sustainable or collapses into a bubble, as trust has been closely linked to technology adoption rates. Yet public opinion polling reveals considerable skepticism about AI. People worry about safety, data privacy, job market impacts, and existential threats. Press headlines regularly warn that AI will rob jobs, that chatbots harm mental health, and that agentic systems run beyond human control.
This skepticism isn't irrational—it reflects the gap between the promises and the reality. When companies claim AI will revolutionize everything while simultaneously laying off workers and failing to generate returns, trust erodes. When the promised productivity gains don't materialize and adoption rates decline among large companies, the narrative begins to crack.
So what does this mean for regular people navigating this moment of economic uncertainty?
First, recognize that the current economic statistics are misleading. If the economy appears strong on paper while you're struggling, you're not imagining things. The growth is concentrated in AI infrastructure spending by a handful of companies, not in wages or opportunities for most workers.
Second, understand that job displacement from AI may be overstated in the near term. Despite the apocalyptic warnings, companies are discovering that AI can't simply replace human workers. The 95% failure rate of AI pilots suggests that the technology isn't ready for the revolutionary transformation being promised. But this doesn't mean complacency—it means the disruption will be messier and more prolonged than advertised.
Third, prepare for volatility. Whether you're an investor with retirement savings tied to tech stocks, a worker in a field supposedly vulnerable to AI automation, or simply a citizen of an economy increasingly dependent on AI hype, the burst will create turbulence. The question is whether we'll use the opportunity it creates to build a more equitable economy or simply watch another transfer of wealth from regular people to those already at the top.
Fourth, pay attention to the capital starvation effect. If you work in manufacturing, healthcare, education, clean energy, or any non-tech sector, the AI bubble is already affecting you through higher costs of capital and reduced investment. When the bubble bursts, it may finally free up resources for innovation and growth in these neglected sectors.
Finally, remember that bubbles don't mean the underlying technology is worthless. The internet was transformative despite the dot-com crash. AI will likely prove useful in specific applications even as the grandiose visions collapse. The challenge is separating signal from noise, legitimate innovation from speculative frenzy.
Former Intel CEO Pat Gelsinger believes the AI bubble will last "several years" before ending, as major disruptive technologies will develop in the latter part of this decade. Others see the correction as imminent. The truth is that timing a bubble's burst is nearly impossible—until it happens, and then it seems obvious in retrospect.
What we can say with certainty is that the current trajectory is unsustainable. You cannot spend $400 billion annually on infrastructure that generates $12 billion in consumer revenue indefinitely. You cannot have seven companies control one-third of the stock market without systemic risk. You cannot starve entire sectors of the economy to feed a speculative frenzy without consequences.
The AI bubble will burst. The question isn't if, but when—and whether we'll be wise enough to use the aftermath to build something better than what came before. The technology will remain, stripped of its most grandiose promises but still useful. The challenge is ensuring that what rises from the ruins serves human flourishing rather than simply setting up the next cycle of speculation and disappointment.
For now, we wait in this strange liminal space where trillion-dollar deals are announced daily, productivity gains remain elusive, and the gap between vision and reality grows ever wider. It's the storytelling economy, and as long as enough people believe the story, the bubble can keep expanding. But stories, no matter how compelling, eventually must confront reality. And when they do, we'll discover what was real beneath all the hype—and who pays the price for the delusion.
