Only if its a bubble.
What happened in the dot-com bubble
The collapse of the internet-era technology boom of the late 1990s remains one of the canonical cautionary tales in financial markets. Sometimes called the Dot‑com Bubble, this era offers important lessons for investors today.
The build-up
Between roughly 1995 and March 2000 many technology companies, especially those with “.com” in their name or closely linked to the growth of the internet were valued on expectations rather than earnings. The index roughly associated with them, the Nasdaq Composite, rose approximately 400% from 1995 to 2000.
According to one academic paper, “many optimistic investors arriving to the market willing to pay high prices for Internet stocks” drove the jump.
Fundamentally, many of these firms lacked business models and sustainable earnings, yet enjoyed astronomical valuations. One summary definition: a “bubble is the act of change in prices which is not supported by analysis of the fundamentals of the asset.”
The burst
By March 2000 the peak was reached, and by October 2002 the Nasdaq had fallen some 78% from its peak.
Many companies were liquidated or were consolidated. The impact extended well beyond the tech sector: ancillary industries (infrastructure, advertising, services) also suffered.
Research shows that part of the problem was speculative behaviour, weak profitability, loose underwriting and optimism unanchored in fundamentals.
Aftermath and lessons
Many investors lost large sums; for example, venture capitalists who were heavily tilted in dot-com stocks.
Yet some survivors emerged (for example, Amazon.com, and eBay) and went on to dominate the digital economy.
A key lesson: valuations must eventually align with sustainable economics (revenues, profits, and cash flows).
Are we in a similar “AI bubble”?
The question posed (“Will the AI Bubble Pop Like the Dot-com Bubble?”) is relevant to many investors today. Below, I examine the similarities, the differences, and what research tells us about the probabilities.
Similarities with the dot-com era
1. Rapid valuation expansion on hopes of fundamental disruption:
A widespread belief that generative AI and adjacent technologies will transform productivity, labour markets, entire industries has created enormous investor enthusiasm. For example, the Bank of England recently warned that “equity market valuations appear stretched, particularly for technology companies focused on artificial intelligence.”
Some analyses argue the current AI valuation surge is actually larger than the dot-com bubble.
2. Exuberant sentiment, narrative-driven valuations:
The sociologist-economist Robert J. Shiller argued in his book Narrative Economics that market valuations are often driven by prevailing stories. In the current climate, much attention is on AI as the next “electricity” or “internet” revolution.
Indeed, one study refers to this as the “AI valuation paradox” where valuations are anchored to imagined future capabilities rather than current earnings.
3. Massive capital flows, speculative vehicles:
Venture-capital funding in AI has surged. For example, in Q1 2025 AI startups raised nearly $73.1 billion globally; about 57.9 % of all global VC funding in that quarter.
Some companies are trading at extremely high revenue multiples (or no revenue at all). For example, my research noted one AI startup, Cognition AI, with merely $300K in revenue, has raised over $500 million, and “one analysis found a $600 billion gap between AI revenue expectations and actual revenue.”
4. Infrastructure and hype spending:
Huge investments in data centres, chips, AI research, cloud infrastructure are reminiscent of build-outs during the internet era. For example, research indicates that only ~5 % of AI adoption projects delivered measurable gains in one study of over 300 projects.
Given these parallels, it is entirely plausible to think that we are seeing a bubble-type phenomenon.
Notable differences
However it is not a perfect mirror of the 2000 era. Some important distinctions:
In the dot-com era, many companies had no viable business models, burned cash and had no path to profitability. Today, many AI-focused companies (especially large tech firms) already generate substantial earnings and own ecosystem advantage. For example, the CEPR paper found that the P/E ratios of the largest tech companies remain below dot-com extremes, even though valuations have surged.
The underlying technology in AI arguably has more substantial real-world applications, though adoption is still uneven. The World Economic Forum observes that, while plenty of hype, there is also significant actual investment in components (data centres, chips) required for AI.
Policy and systemic responses are fundamentally different. This is a notable difference. After the 2008 crisis, central banks adopted unprecedented tools like Quantitive Easing, and Liquidity Backstops, which were less mature in 2000s. This may blunt or alter the nature of a crash scenario given the powerful elixir behind tools like Quantitative easing.
Regulation, investor sophistication, disclosure norms have greatly evolved since the 2000s.
The market is more global and diversified today: AI infrastructure is deployed across many sectors and geographies.
What research and signals tell us
Here are some of the key findings and warning signs:
Valuation metrics are elevated. Analysts find that certain valuation multiples (P/E, P/S) for tech companies are “approaching the levels observed during the dot-com era.”
Some firms show stretched expectations: as the Guardian notes, “valuations are extremely stretched by usual standards.”
For many AI projects, the realised return is disappointing: one MIT-linked study found only ~5 % of AI projects delivered measurable gains.
Scholars propose a “Capability Realisation Rate” (CRR) model which quantifies the gap between what the market expects from AI and what is being delivered.
The World Economic Forum cautions that while influence from AI is real, the “bubble” may be partially in the valuation of companies, not necessarily in the utility of AI itself.
Some voices are outright: for example, one Seeking Alpha opinion piece declares “the AI bubble appears to be teetering on the brink of bursting.”
Institutions of importance are also sounding alarms. For instance, the Bank of England flagged stretched valuations and risk of sudden correction.
Will it pop like the dot-com bubble?
Or: “Will it wipe out 80% of its value?” is a provocative one. Here is my view:
- An 80 % collapse across the entire tech and AI segment is less likely than in the dot-com case, but it is not impossible.
- A full-blown 80 % drop is less likely because many of the large AI companies are already profitable, entrenched, diversified, and have substantial balance sheets. Unlike many dot-com companies of 1999, the modern day AI companies are not purely speculative.
- Because central banks and governments are now more equipped to intervene if systemic risk emerges.
- The use-case story for AI arguably has more substance (though it remains to be widely realised), which means the technology is not just hype in the way many dot-com firms were.
- The market is broader and more globalised, meaning risk is more distributed.
Why a significant correction is still a real risk
If the productivity gains, revenue growth and business-model realisations fall well short of expectations, valuations could drop sharply.
If many AI projects fail to scale, or structural bottlenecks like power, data, and chip supply emerge, there are warning signs on that front.
If investor sentiment turns, narrative-driven valuations can unwind rapidly. Research on bubbles shows herding, momentum trading, and optimism cycles.
Smaller “AI-labelled” companies may be extremely overvalued and vulnerable; while the big ones may survive, many minnows could collapse.
No market crash happens uniformly; even if the largest firms hold up, valuations could contract by 30-50% or more in the sector.
In short, I judge an all-out 80 % wipe-out across the big tech and AI segment to be unlikely, but I would certainly regard a double-digit percentage correction (say 20-30 %) as plausible if expectations mis-align and adoption lags.
Takeaways
1. Quantitative Easing and central bank tools matter:
Systemic risk may be mitigated compared to 2000 due to the tried and tested, powerful tools at the disposal of Central banks in times of crisis. However, it also implies that markets may become more reliant on monetary largesse, which tends to be a moral-hazard issue.
2. Narrative-driven valuations are more dominant now:
The story around AI is driving valuations. But that also means when the narrative fades, valuations can tumble. Hype-driven markets have limited tolerance for disappointment.
3. Diversification remains wise:
Since the risk is asymmetric, even if the most prominent AI companies avoid collapse, many smaller firms may suffer. And even the large firms could face earnings disappointments that dent valuations. Therefore, diversification across asset classes, geographies, and sectors remains important.
4. Don’t assume identical mechanics to 2000:
While the parallels are strong, the context is different in key metrics like profitability, large-cap strength, macro tools, and globalisation. Therefore designing your risk-management based on a perfect 80% collapse scenario may be misled. The correction might instead be more moderate or manifest differently. For instance, a multi-year period of under-performance rather than a single crash.
Time horizon matters: If you are a long-term investor (10+ years) you might tolerate more volatility and see AI as structural change. If you have a shorter horizon, you might consider trimming exposure or increasing hedges.
Stay alert to signals: Key red-flags: flattening revenue growth despite high capex, major cost-overruns in AI infrastructure, key regulatory headwinds, and media narrative turning negative.
Avoid trying to time the peak: Bubbles often “end” with a melt-up (everyone jumps in) before the correction. Recognising the top is extremely difficult so its better to control risk proactively rather than reactively.
Concluding Remarks
We very likely are in a geo-political and technological boom with bubble-like attributes (exuberant valuations, narrative dominance, heavy capital flows). However, whether it pops like the dot-com bubble (−80 % across the board) is far less certain because the structure today is fundamentally different; the large players are more robust, and policy tools are far more entrenched.
What is far more plausible is a significant correction perhaps a drop of 30-50% in valuations for many AI-labelled companies, or a multi-year period of under-performance for the sector. As such, prudence suggests that investors should not assume everything will continue to rise unchecked, and should take risk-mitigation measures accordingly: diversification, selective exposure, and monitoring fundamentals.
In short, the “bubble” may not explode in a dramatic single event, but it may deflate or shift in ways that cause meaningful losses for those over-exposed. Being prepared is more important than being certain of the timing.
Sources: Guardian, ReutersReuters, AP NewsAP News, Yahoo Finance, and Bloomberg.