Is AI the Next Dot-Com Bubble? If AI Crashes, What Happens to India?

Introduction: The Lessons of Technological Revolutions

Imagine you are in the year 1999. You meet an investor who tells you that within the next 20 years, people would buy products online, watch movies online, communicate through social media, store data in the cloud, and carry the internet in their pockets. It would sound unbelievable, isn’t it? Now imagine that same investor also tells you that the internet stocks are about to lose 80% to 90% of their value over the next couple of years. Would you believe that?

The surprising part is that both predictions turned out to be true. The internet transformed the world, but many investors still lost massive amounts of money during the dot-com crash. And the internet is not the only example. Throughout history, new technologies have repeatedly created periods of extreme excitement. Railways, electricity, automobiles, telecom networks, and now the internet all triggered massive investment booms. In many cases, the technology eventually succeeded, society benefited, and businesses changed forever, but shareholders often suffered because expectations became too optimistic and capital flooded into the sector far faster than the eventual profits justified.

There’s a fascinating book called Engines That Move the Markets that studies more than two centuries of technological revolutions. It says technologies change, but human behavior does not. Investors repeatedly make the same mistake: they overestimate short-term growth, underestimate competition, and confuse a great invention with a great investment. Today, there is one technology that is dominating the global market in exactly the same way: artificial intelligence.

The Scale of the AI Investment Boom

To put the scale into perspective, the four largest hyperscalers in the world—Google, Amazon, Microsoft, and Meta—are expected to spend nearly $600 billion on AI-related capital expenditure in 2026 alone. At the same time, research firms like Gartner estimate that worldwide AI spending could reach approximately $2.6 trillion in 2026.

And that is only the beginning. According to a report by McKinsey & Company, cumulative global investment in data center infrastructure could approximate $7 trillion by 2030. In other words, companies are not just spending money on AI software; they’re simultaneously building the physical infrastructure like data centers, chips, power systems, and networks required to support the AI economy.

Now, let me be very clear: there’s no denying that AI could become one of the most transformative technologies in our lifetime. The question is not whether AI will change the world; the question is whether investors and corporations are becoming too optimistic about how quickly AI can generate economic value. History teaches us a very important lesson: sometimes the technology is real, the adoption is real, and the impact is real, yet investors still lose money. That usually happens when people stop distinguishing between a great technology and a great investment.

This is exactly why investors today are making comparisons between the AI boom and the internet boom of the late 1990s. Back then, investors correctly predicted that internet usage would explode, and internet traffic did grow exponentially. However, telecom companies and network operators started building infrastructure at an even faster pace. Billions of dollars were invested in laying fiber optic cables across countries and oceans. The problem was not weak demand; the problem was excess supply. Capacity expanded much faster than actual usage. Eventually, prices collapsed, returns collapsed, and many infrastructure companies destroyed shareholder wealth despite being right about the long-term future of the internet.

Fast forward to today: every major technology company is racing to build AI capabilities. Every venture capitalist wants exposure to AI, institutional investors are chasing AI, retail investors are chasing AI, and technology companies are investing hundreds of billions of dollars into chips, data centers, LLM models, and infrastructure. This is where the debate around an AI bubble begins. Are we witnessing the early stages of a technology revolution that will justify these investments, or are we witnessing another period where expectations and capital spending have moved far ahead of economic reality?

The Bull Case: Why the AI Boom is Justified

Let’s first look at the arguments supporting the AI boom and why many experts believe that current AI investments are completely justified.

• Fundamental Profitability: The first argument supporting the fact that this AI boom is not like the dot-com era is that back then, many internet companies had no profits, no sustainable business model, and in some cases, barely any revenue. Investors were simply buying dreams. Today, companies leading the AI race—like Microsoft, Google, Amazon, Meta, and Nvidia—are some of the most profitable businesses in human history. They’re not raising money to survive; they are using their own cash flow to fund AI investments.

• Enormous Cross-Industry Opportunity: The second bullish argument is that the opportunity itself is enormous. AI could potentially impact healthcare, banking, manufacturing, education, logistics, legal services, and almost every knowledge-based industry. In other words, AI is being viewed as a general-purpose technology, similar to electricity or the internet. If that turns out to be true, then today’s spending may look expensive but not irrational.

• Accelerating Corporate Demand: The third argument is perhaps the most important one. The biggest technology companies are not slowing down; in fact, they are accelerating their investments. Every major hyperscaler is increasing AI spending because nobody wants to be left behind. Alphabet CEO Sundar Pichai recently admitted that even Google is struggling to keep up with the demand for AI products and services. That’s a very important point: most CEOs are not saying demand is weak; they’re saying demand is stronger than expected.

• Institutional-Led Structure: The fourth argument is that the structure of this AI boom is very different from 1999. The dot-com boom was largely driven by retail investors chasing anything with a “.com” attached to its name. Today’s AI boom is being driven primarily by institutional pension funds, sovereign wealth funds, and large asset managers. That does not mean prices cannot fall—they absolutely can—but it makes the price action less panicky than in 1999.

• Physical Infrastructure Scarcity: Finally, there’s a very interesting argument around power and infrastructure. Today, one of the biggest bottlenecks in AI is not demand; it is electricity and data center capacity. Building large AI infrastructure takes years, power approvals take years, and data centers take years. As a result, existing infrastructure has become extremely valuable. Some analysts actually argue that this scarcity could protect the profitability of current AI infrastructure providers because new supply cannot be created overnight.

The Bear Case: Signals of an Entering Bubble

Now, let’s look at the other side of the argument and examine the concerns that AI may already be entering bubble territory.

• The Revenue-to-Capex Disconnect: The first concern is that AI investments are growing much faster than AI revenue. Today, companies are spending hundreds of billions of dollars building data centers, buying GPUs, and securing power infrastructure. The question is simple: will these investments generate enough revenue to justify the spending? Some analysts believe the gap is becoming too large. This is very similar to what happened during the internet boom, where investors underestimated how long it would take for infrastructure investments to become profitable.

• Elevated Market Valuations: The second concern is valuation. There are a few indicators that investors use to judge whether the overall market has become expensive. One of them is the Shiller PE ratio (CAPE ratio). Instead of looking at a single year’s earnings, it compares stock prices with inflation-adjusted average earnings over the last 10 years. Historically, the US market has traded around 17 to 20 times on this metric. During the dot-com peak in 2000, the ratio crossed 44 times, and recently it crossed 40 times again. While not exactly at dot-com levels, it is close enough to make people nervous.

• The Buffett Indicator and Cash Reserves: Another popular indicator is the market cap-to-GDP ratio, often called the “Buffett Indicator.” During the dot-com bubble, the total US stock market was roughly around 140% of US GDP. Today, that number sits around 230%. In other words, the stock market has become significantly larger relative to the size of the economy than it was even during the internet mania.

  Furthermore, a very interesting signal comes from Warren Buffett himself. While everyone is talking about AI and chasing growth stocks, Berkshire Hathaway has quietly accumulated one of the largest cash positions in history, sitting on hundreds of billions of dollars in cash and US Treasury bills. Warren Buffett has not explicitly said that AI is a bubble, but historically, whenever he struggles to find attractive opportunities, he prefers holding cash rather than overpaying for assets. One of the greatest investors of all time is effectively saying, “I would rather earn 4% to 5% in Treasury bills than buy many stocks at current valuations.”

• Historic Market Concentration: The third concern is market concentration. In fact, this may be one area where today’s market is even more extreme than in 2000. The “Magnificent 7” stocks compose more than one-third of the total market cap of the S&P 500. This concentration creates fragility. When Cisco tanked in 2000, it sank every broker’s portfolio. Today, any significant correction in Nvidia or Google would ripple through virtually every passive investment vehicle on this planet. If you observe the NASDAQ, it continues to make new highs, but the rally has been heavily concentrated in a few AI giants; otherwise, the overall market breadth in the S&P 500 is quite weak.

• Debt-Financed Infrastructure: The fourth concern is debt. During the dot-com era, most speculation was funded through equity. Today, a growing portion of AI infrastructure is being financed through the debt market. If AI investments fail to generate expected returns, the risk could extend beyond shareholders and ripple into the broader credit system.

• The Adoption Lag: Finally, there’s the adoption lag problem. Almost everyone agrees that AI will eventually transform industries, but the debate is about timing. Investors are valuing companies based on the assumption that AI monetization will happen quickly. History suggests that technological revolutions often take much longer to generate profits than the market expects.

What Happens Next? Navigating the Uncertainty

After looking at both sides of the argument, the obvious question is: what happens next? The truth is, nobody knows with 100% certainty. In fact, one of the biggest mistakes investors make during a bubble is assuming that because something looks expensive today, it must fall tomorrow. History does not work like that. The NASDAQ looked expensive in 1997, it looked even more expensive in 1998, and it looked absolutely absurd in terms of valuation in 1999, yet it kept going higher before finally collapsing in March 2000. Investors who correctly identified the bubble but exited too early missed one of the biggest rallies in market history.

The same thing can happen with AI. Even if someone believes AI valuations are stretched, that does not automatically mean a correction is around the corner. The reality is that AI spending is still accelerating, hyperscalers are increasing capex, demand for GPUs continues to exceed supply, and enterprise adoption is still in its early stages.

The key variable to watch is not stock price movement, but economics. The moment investors start seeing signs that capex is rising much faster than actual monetization, sentiment can change very quickly. If we think about the next few years, there are broadly three possible outcomes:

1. The Bullish Outcome: AI revenues continue growing rapidly, enterprises figure out how to monetize AI faster than expected, and productivity gains start showing up across industries. In this scenario, markets may witness normal corrections of 10% to 20%, but the broader AI uptrend remains intact.

2. The Realistic Cooling Scenario: AI succeeds, but expectations cool down. If revenue growth takes longer than expected, or if investors simply decide that current valuations are too expensive, a 25% to 40% correction in AI-related stocks becomes entirely possible. Personally, this is the scenario I find most plausible.

3. The Extreme Dot-Com Style Bust: This extreme outcome would require multiple things to go wrong simultaneously: AI adoption disappoints, economic growth weakens, capex slows dramatically, and investors start questioning the economics of the entire AI ecosystem. In that situation, AI-related stocks could potentially witness drawdowns of 50% to 60% or even more.

However, the probability of a 70% to 80% NASDAQ collapse similar to 2000 is significantly lower today because the companies leading the cycle are fundamentally different. Back then, companies had no real business models. Today, Nvidia, Microsoft, Alphabet, Meta, and Amazon generate tens of billions of dollars in free cash flow and profit. Even if valuations compress, the underlying businesses remain strong.

The Impact on India: Vulnerability or Strength?

If the US market corrects because of an AI reset, India will not be immune in the short term. India will almost certainly feel the pain: foreign investors could reduce risk across emerging markets, technology stocks may correct, market sentiment could weaken, and sectors linked to the AI infrastructure buildout—including data center themes and AI capex beneficiaries—could face a massive correction.

But here is where the Indian story becomes interesting. Unlike Taiwan or South Korea, India is not at the center of this AI trade. Taiwan’s market is heavily dependent on semiconductor manufacturing, and South Korea is deeply exposed to memory chips. India is different; in fact, India has been one of the biggest laggards in this AI race, and that could actually become a strength if the AI trade starts unwinding. If global investors begin rotating away from AI-heavy markets, they will eventually look for economies with stronger domestic growth drivers, and India fits that description quite well.

At the end of the day, India’s growth story is still driven by consumption, banking, housing, infrastructure, manufacturing, and financial services. AI may influence these sectors, but it is not the primary engine behind them. This is why I believe India could initially fall along with the rest of the world but potentially recover faster than many AI-heavy markets.

There is also another interesting angle regarding the Indian IT sector. One of the biggest fears surrounding Indian IT is that AI will eliminate software jobs and permanently weaken the outsourcing model. But what if AI deployment turns out to be slower, more expensive, and more complicated than the market currently expects? Then, many of the disruption fears built into Indian IT valuations may start fading. Global enterprises would still need cloud migration, cybersecurity, data engineering, AI integration, governance, and workload redesign. If AI becomes an implementation challenge rather than simply a coding challenge, Indian IT companies could actually become beneficiaries, allowing India to emerge as one of the most resilient markets globally.

Conclusion: Preparing for Multiple Futures

In the current environment, a sensible approach is to remain diversified across equities, gold, and fixed-income instruments. Equities remain the best asset class for long-term wealth creation, and completely avoiding them because of bubble fears can lead to significant opportunity loss.

At the same time, keeping a reasonable allocation to debt or fixed-income instruments provides stability and, more importantly, liquidity. If the market witnesses a meaningful correction, investors with some cash allocation will have the ability to deploy capital at much more attractive valuations. Gold also deserves a place in the portfolio because it acts as a hedge during periods of geopolitical uncertainty, currency weakness, inflation concerns, and financial stress.

Ultimately, investing is not about predicting a single future; it is about preparing for multiple futures. Maybe AI continues driving markets higher, maybe there’s a 30% to 40% correction, or maybe geopolitical risks intensify—nobody knows. A diversified portfolio may not be the most exciting strategy during a bull market, but it is often the most effective strategy for navigating periods of uncertainty. In investing, staying in the game is usually far more important than trying to predict perfectly what happens next.

Source: Is AI the Next Dot-Com Bubble? If AI Crashes, What Happens to India?