AI Bubble: Will It Pop Like Dot-Com—and When?
Lessons from 2000 on valuations, debt, and crowded trades in today's AI frenzy
In this post, we would like to compare what is happening with the AI industry with the 2000 dot‑com bubble.
We do not think the question to focus on is “Is AI a bubble?”—bubbles burst no matter what we call them. The art here is being able to guess the impact of the explosion. Some bubbles burst and nobody even notices, and some bubble bursts and the entire global system stops functioning. AI won’t vanish overnight; at worst, it might hit the brakes.
In our view, the most important part of any financial bubble is the timing—and actually, that’s easier to answer than it seems: it pops exactly when liquidity in the markets dries up or disappears. With upcoming quantitative easing keeping the taps open for the foreseeable future, we don’t see that happening anytime soon; instead, the real wildcard may be how long investors stay patient financing a sector that still isn’t generating enough revenue to justify the hype. So our question here isn’t simply “Is it a bubble?” but rather: could it pop anyway, and if yes, when and how big will the impact on the rest of the economy be? As professionals who lived through the 2000 dot‑com era as traders, analysts, and investors, a lot of what we see today feels familiar—but not identical.
Echoes of the Dot-Com Hype: Big Stories, Bigger Valuations
The current AI boom shares many of the classic ingredients of the dot‑com bubble: big stories, big valuations, and heavy debt exposure behind the scenes. Back then, “the internet will change everything” was enough to float companies with barely any revenue; today, “AI will transform every industry” plays a similar role, stretching price‑to‑sales multiples and embedding assumptions of near‑flawless execution. The key twist this time is that much of the risk is concentrated in profitable giants like Microsoft, Alphabet, Amazon, and Nvidia, which really do have strong cash flows and growing AI businesses—but still nothing close to the earnings implied by the money being poured into the theme.
Scale and Concentration: A Narrower, Riskier Bet
The scale and concentration of the current boom also set it apart. The dot‑com frenzy was spread across hundreds of small and mid‑cap IPOs, many of which disappeared when the music stopped; AI, by contrast, is dominated by a small cluster of mega‑caps and a focused ecosystem of hardware, foundation models, and “AI platform” players. For investors, that means the upside has felt very smooth and index‑driven—but it also means that if sentiment turns, almost everyone is exposed to the same few names at the same time.
Overinvestment: History Rhymes with Fiber and Chips
On the investment side, the rhyme with 2000 is hard to ignore. Then, telecoms and internet firms massively overbuilt fiber and data centers based on heroic traffic forecasts and struggled under the debt when reality disappointed; now, companies are committing extraordinary sums to AI data centers and chips with an equally ambitious story about future demand. In both cases, excessive borrowing quietly sits underneath the narrative: long‑dated capex commitments, balance‑sheet risk, and structured bets tied to a narrow group of “winners” that can turn a healthy correction into something much more painful.
Stronger Fundamentals? Adoption Gives AI an Edge
There are real differences, though, in adoption and staying power. Unlike many dot‑com darlings, which went public with little more than a business plan and a URL, today’s AI players are selling tools that are already embedded in workflows at most large organizations, and core suppliers are reporting very real AI‑related revenue growth. That is why an eventual AI “burst” is more likely to look like a long, grinding de‑rating—multiples compressing, expectations coming down, some high‑profile blow‑ups—rather than a total wipeout of the whole space.
The Real Danger: Extreme Concentration in Holdings
What worries us most is not just overvaluation and excessive debt, but extreme concentration in holdings. When the same handful of AI stocks sit at the heart of index funds, active funds, hedge funds, and retail portfolios, disappointment in one corner of the story can quickly turn into forced selling across the board. Overvaluation sets the stage, heavy debt provides the accelerant, and crowded positions narrow the exit. If and when this cycle turns, the experience may feel different from 2000 in its details—but for investors, asset allocators, and policymakers, the core challenge will be the same: navigating the messy period when an important technology keeps advancing, even as the financial promises built on top of it are being rapidly repriced.
This newsletter is for informational and educational purposes only—not financial advice. Use of this information is at your own risk, and the authors are not responsible for any resulting losses or damages.