Are We Already Living in an AI Bubble?
By JT Mercer — 2025-11-28
Investors cannot decide whether the current artificial intelligence boom is a once-in-a-generation opportunity or the prelude to an AI bubble that ends the way the dot-com frenzy did. In the same week that chipmakers and cloud giants posted record earnings, central bankers, star fund managers, and household-name economists lined up to warn that the AI trade is starting to look dangerously familiar. The question is not whether there is hype; the question is whether this AI bubble story matches the numbers on the ground or just the fear in people’s heads.
What Happened
Start with the scoreboard. In 2025, the S&P 500’s roughly 15% gain has been driven by a handful of tech names whose fortunes are tied directly to AI infrastructure and software. The so-called “Magnificent 7” now account for about 37% of the index’s total market value, a concentration that rivals the late-1990s when the world convinced itself that any company with a “.com” domain deserved a premium. CBS News
Nvidia has become the poster child for the AI trade. In its latest quarter, the company reported around $57 billion in sales in just three months, part of a fiscal year in which revenue more than doubled to about $130 billion and profit surged roughly 145%. CBS News+1 Those figures are not the revenue-free fantasies of 1999; they are concrete earnings tied to real demand for compute. Yet the same earnings report that sent the stock higher also reignited debate about an AI bubble, because every blowout quarter raises the stakes for what happens when demand inevitably slows. Investopedia+1
The fear is not just retail investors spooked by red charts. The European Central Bank’s latest Financial Stability Review singled out US tech valuations as “stretched,” driven by investor FOMO around AI and a renewed risk-on mood after April’s sell-off. The ECB stopped short of declaring an AI bubble, but warned that a sharp correction is plausible if today’s rosy assumptions about AI adoption and profitability fail to materialize. Financial Times
On the asset-management side, GMO, the firm co-founded by long-time bubble-watcher Jeremy Grantham, described the AI boom as a “classic investment bubble” in its November letter. The note pointed to eye-watering rallies not just in the headline AI chipmakers, but in more speculative corners like quantum computing stocks, where some names have climbed over 1,000% on promises, not profits. Business Insider
Then there is Michael Burry, still trading on his “Big Short” reputation. He has disclosed put positions against Nvidia and Palantir, arguing that AI infrastructure is being depreciated too slowly and could face painful writedowns if hardware becomes obsolete faster than accounting schedules admit. He explicitly frames this as part of a broader AI bubble in which investors underestimate just how quickly today’s state-of-the-art data centers could look antique by 2028. Business Insider+1
Survey data show that this is not just a talking-head obsession. Bank of America’s October fund manager survey found that 54% of respondents already view AI-related assets as being in bubble territory. Analytics India Magazine When Wall Street sold off sharply last week, headlines in London and New York alike pinned the move on a resurgence of AI bubble fears rather than on any single economic data point. The Guardian+1
At the same time, defenders of the trade – including BlackRock’s iShares unit – argue that while AI valuations are high, they are nowhere near the extremes of 2000. Back then, top tech leaders traded around 70 times two-year forward earnings; today the core AI hyperscalers – Microsoft, Alphabet, Amazon, Meta – sit closer to 26 times. BlackRock Brookings and Yale analysts have published their own versions of the same uneasy verdict: there are clear risk indicators, but the evidence that an AI bubble has already crossed the line from exuberance into insanity is mixed. Brookings+1
In short, both sides have ammunition. The market has the feel of an AI bubble, but the income statements are still putting up numbers that would have stunned anyone staring at Pets.com in 1999.
Why It Matters
This is not some niche argument among hedge funds. If you own a target-date retirement fund, odds are high that you are long the AI trade whether you like it or not. With roughly a third of the S&P 500 tied up in a short list of tech-driven names, any disorderly unwinding of an AI bubble would not stay inside Silicon Valley. It would flow straight through to 401(k) balances, endowments, and public pension funds that benchmark to broad indices.
The real fear is not that Nvidia misses a quarter, or that one mega-cap trips over a bad product launch. The fear is that the entire AI narrative breaks at once. Companies have already issued more than $140 billion in AI-related corporate debt this year, surpassing all of 2024’s issuance, to fund data centers, power infrastructure, and chip purchases. IR Impact If it turns out that customers cannot fully monetize the AI services built on top of that hardware, the write-offs and capex reversals will not be confined to a single ticker.
The Systems Angle
Look under the hood and the AI trade starts to look less like a simple growth story and more like a tightly coupled system. As Yale’s analysis points out, Microsoft accounts for nearly a fifth of Nvidia’s revenue on an annualized basis. At the same time, Microsoft is a major shareholder in OpenAI, which in turn is a key buyer of GPUs and a driver of AI workloads that flow through Microsoft’s cloud. CoreWeave, another big AI cloud player, counts both Nvidia and Microsoft as investors, while also depending on their chips and capital. Yale Insights
This is not inherently fraudulent, but it is circular. Capital flows from one balance sheet into another, back into equity stakes, and around again into customer contracts. In a boom, this kind of circularity amplifies returns. In a bust, it can amplify losses. That is the scenario AI bubble worriers have in mind: not a clean, contained correction, but a multi-node feedback loop where one failed expectation cascades through suppliers, customers, and financiers at once.
The Information War
Fear of an AI bubble is also a narrative asset. For politicians worried about speculative excess and regulators wary of financial stability, talking down AI euphoria is an easy way to look prudent. For active fund managers who underweighted tech and underperformed, warning of an impending crash is a way to defend their positioning while they wait for vindication. For AI-heavy executives, dismissing AI bubble talk as hysteria is an equally predictable move.
Google’s Sundar Pichai has acknowledged “elements of irrationality” in AI investing even as he insists the technology will touch every company. In contrast, Alibaba’s leadership has waved off AI bubble fears and doubled down on investment plans despite margin pressure. The Times of India These are not neutral signals; they are part of the information war over who controls the pace and framing of the AI boom.
Media dynamics are not helping. As the World Economic Forum has noted, spikes in “AI bubble” chatter track almost perfectly with headline events: a big earnings beat, a high-profile AI outage, a regulatory hearing. World Economic Forum Every surge in the phrase AI bubble feeds investor anxiety, which in turn makes markets more jumpy. The fear of being the last one holding the bag becomes its own fundamental.
From the Book
If you read Book 1, “The Unmaking of America,” this entire AI bubble debate may feel familiar, even though the novel is about politics, not stock tickers. Late in the campaign, Vera2’s “Where Is My Ballot?” field report documented how a handful of decision points and system quirks could tilt an entire state’s election without most voters ever seeing the levers move. The point of that chapter was simple: concentrated power plus opaque systems equals fragility, even when everyone involved insists the system is “resilient by design.”
We are watching a similar pattern play out in the AI economy. A small number of hyperscalers, chip designers, and model labs dominate the infrastructure. Their balance sheets, customer contracts, and capital spending plans cross and recross in ways that make the whole machine look robust – until you ask what happens if two or three of those assumptions break at once.
Book 1 also hammered on narrative capture. Official talking points about “isolated glitches” held up right until Vera2 surfaced a pattern that the public could no longer ignore. An AI bubble, if it exists, will end the same way. The trigger will not be a hot take or a gloomy editorial. It will be a run of concrete data: canceled data-center projects, slowing AI usage metrics, and CFOs quietly trimming AI budgets because the promised productivity gains are not showing up.
What We Don’t Know Yet
Despite the confident language on both sides, no one can say with certainty whether we are in an AI bubble until after the fact. Brookings has suggested a set of indicators to watch: the pace of AI investment relative to profits, the timelines for data-center construction, real adoption levels in core industries, pricing power for AI services, the degree of competition, and public trust in the technology. Brookings Those dials are moving, but many of them still point to “overheated but not absurd.”
We also do not know how quickly AI will filter into the unglamorous parts of the economy: logistics, insurance underwriting, clinical documentation. If those use cases scale, earnings may grow into the valuations. If they stall, the AI bubble narrative will have teeth, and fast.
Counterarguments & Rebuttals
AI bubble skeptics make three main points. First, fundamentals exist this time. Nvidia, Microsoft, Alphabet, and their peers are printing cash rather than selling page views and pets’ chew toys. Second, valuations, while rich, are nowhere near the extremes of early-2000 tech. Third, much of the AI build-out is funded from profits and strong balance sheets, not debt, which reduces systemic risk even if a correction hits. BlackRock+1
All of that is true as far as it goes. A market driven by an AI bubble with no earnings to back it would already have collapsed. Instead, we have companies whose income statements still justify substantial multiples. That is exactly why the fear is so jittery: nobody wants to be early in calling the top on a trade that keeps rewarding believers.
The rebuttal is that bubbles usually end from the margin, not the core. You do not need Nvidia to implode to have an AI bubble that hurts. You need a few high-profile flame-outs, some ugly downgrades in the AI “adjacent” names, and a shift in corporate behavior from “build at any cost” to “prove the ROI.” Deutsche Bank’s research on an earlier downdraft in AI-linked stocks framed it as the “AI bubble” bubble bursting – a crack, not a shatter – and reminded readers that these processes often unfold in stages. DB Research
In other words, the presence of real earnings makes a crash less likely, but it does not immunize the market against a long, grinding deflation of AI bubble enthusiasm.
What to Watch Next
If you want to know whether the AI bubble fear is justified, stop staring at the daily share-price swings and track three things.
Watch capex guidance from the hyperscalers and AI chip buyers. Do they keep escalating their data-center budgets, or do we see the first signs of plateauing or delay? Watch enterprise adoption metrics and pricing in earnings calls. Are customers rolling pilots into production at scale, or are they quietly cutting back after low-yield experiments? And watch credit markets. The $141 billion in AI-linked debt issued so far this year is sustainable as long as cash flows arrive on schedule; if spreads widen and issuance slows, that is your first hard signal that lenders are starting to doubt the story. IR Impact+1
If those dials stay green, the AI bubble talk will look like another round of premature doom-saying. If they turn red together, you will not need anyone on television to tell you what is happening.
Key Data Points
So far in 2025, roughly 35–37% of the S&P 500’s value has rested on a handful of tech names, while the index as a whole is up about 15%, a sign of just how concentrated the AI trade has become. CBS News+1 Nvidia’s revenue has surged past $130 billion with profits up more than 140% year on year, yet prominent investors like Michael Burry have taken sizable short positions, betting that accounting and obsolescence risks are underpriced. CBS News+1 Bank of America’s survey shows that 54% of professional investors already describe AI assets as a bubble, and companies have floated about $141 billion in AI-linked debt this year, topping last year’s total issuance. Analytics India Magazine+1 Meanwhile, policy institutions from the ECB to Brookings are explicitly warning about “stretched” valuations and calling for closer monitoring of AI bubble risk, even as big asset managers insist “this time is different” because earnings and balance sheets are stronger than in 2000. Financial Times+2Brookings+2
Further Reading
CBS News — “Artificial intelligence has fired the stock market to record highs. Should you worry about an AI bubble?”
https://www.cbsnews.com/news/artificial-intelligence-ai-bubble-stock-market-economy-dotcom/
Brookings Institution — “Is there an AI bubble?”
https://www.brookings.edu/articles/is-there-an-ai-bubble/
BlackRock iShares — “Are AI Stocks in a Bubble? Why This Isn’t a Dot-Com Redux”
https://www.ishares.com/us/insights/ai-stocks-bubble-2025-valuation-outlook
Yale Insights — “This Is How the AI Bubble Bursts”
https://insights.som.yale.edu/insights/this-is-how-the-ai-bubble-bursts
Business Insider — “GMO warns AI is a ‘classic investment bubble.’ Here’s what to buy instead.”
https://www.businessinsider.com/stock-market-crash-bubble-non-ai-picks-where-to-invest-2025-11
The Guardian — “AI bubble fears return as Wall Street falls back from short-lived rally”
https://www.theguardian.com/business/2025/nov/20/stock-markets-ai-nvidia
Connect with the Author
Curious about the inspiration behind The Unmaking of America or want to follow the latest news and insights from J.T. Mercer? Dive deeper and stay connected through the links below—then explore Vera2 for sharp, timely reporting.
About the Author
Discover more about J.T. Mercer’s background, writing journey, and the real-world events that inspired The Unmaking of America. Learn what drives the storytelling and how this trilogy came to life.
[Learn more about J.T. Mercer]
NRP Dispatch Blog
Stay informed with the NRP Dispatch blog, where you’ll find author updates, behind-the-scenes commentary, and thought-provoking articles on current events, democracy, and the writing process.
[Read the NRP Dispatch]
Vera2 — News & Analysis
Looking for the latest reporting, explainers, and investigative pieces? Visit Vera2, North River Publications’ news and analysis hub. Vera2 covers politics, civil society, global affairs, courts, technology, and more—curated with context and built for readers who want clarity over noise.
[Explore Vera2]
Whether you’re interested in the creative process, want to engage with fellow readers, or simply want the latest updates, these resources are the best way to stay in touch with the world of The Unmaking of America—and with the broader news ecosystem at Vera2.
Free Chapter
Begin reading The Unmaking of America today and experience a story that asks: What remains when the rules are gone, and who will stand up when it matters most? Join the Fall of America mailing list below to receive the first chapter of The Unmaking of America for free and stay connected for updates, bonus material, and author news.
