The AI Boom: Is It a Bubble, or Something Bigger?

Hey guys, I hope you are all doing well. It has been a while since I wrote about a business-related topic. Last time I covered Japan's lost decade, and this time we will be talking about the AI boom, its possible causes, the evidence behind it, and what the consequences might look like. Along the way, we will also draw comparisons to one of the most well-known speculative episodes in market history: the Dot-Com bubble of 2000.

This blog has three parts:

• The evidence and current data

• A comparison with the Dot-Com era

• The possible consequences

This blog has significantly drawn from our article written for FINE 442 (Capital Markets and Institutions) at McGill University's Desautels Faculty of Management, and cannot be regarded as solely personal work but as a team outcome. Co-authors: Mathis Bélanger, Maxime Gauthier, and Tristan Héroux. Full report available:HERE.

The evidence: what the data actually shows

There seems to be a belief among people that an AI crash might occur soon because most of the AI firms are still unprofitable despite having invested heavily in infrastructure. In this section, we shall examine some key metrics – price-to-earnings ratio, profitability and expenditure – to analyze the situation.

Tech firms have started to reach gargantuan valuations. Nvidia has crossed the $3 trillion mark. Large tech firms like Meta, Microsoft, Alphabet, Amazon, and Oracle planned to spend around $342 billion for their capital expenditures in 2025, which marks a jump of 62% compared to last year. This expenditure will be focused on the AI chipsets, data centers and infrastructure. Additionally, the AI stocks have been performing exceedingly well compared to the overall market. However, what is the justification for the same?

At present, the Price Earnings ratio for the S&P 500 stands at about 28-29, higher than its average during the past ten years, which stood at 19.83x. Moreover, the picture becomes quite revealing when the technology sector's Price Earnings ratio is analyzed. The Price Earnings ratio for the technology sector hit a high of 46.2 times in 2025, more than double the Price Earnings ratio for the broad market index. In other words, it clearly indicates that the market does not value today's earnings but also those which would be earned by these companies in the future. Apart from this, in 2025, investments in artificial intelligence start-ups alone accounted for over 50 percent of total venture capital investment globally, which once again shows the popularity of ai and its concentration in the market.

Data from FINE442 Presentation

Dot-Com and AI: similar story, different fundamentals

Now let us discuss the Dot-Com bubble for a brief moment and compare that with the present situation. From January 1995 till March 2000, the value of NASDAQ increased by 600%. This was because people were extremely optimistic about internet-based firms. There was an increase in venture capital from $7 billion in 1995 to almost $100 billion in the year 2000, which is a 14-fold increase in only five years, where internet firms used up almost 80% of all venture capital by 1999. By 1999, as many as 476 firms went public in the United States, which were mostly internet start-ups making no profits.

If we examine the current AI firms, the story becomes quite different. The tech giants behind the AI firmscurrently experiencing massive growth are actually making a profit and have good financial strength. For example, Nvidia has created over $60 billion in free cash flow in 2024, which was definitely not seen with the big players during the Dot-Com era. In fact, Cisco, one of the most favored network infrastructure firms during that time, had an incredibly high price-to-earnings ratio of almost 200 times. Although Nvidia has a high P/E too, this number cannot come close to the ratio mentioned above. One more similarity between the two eras can be identified by referring to market concentration. Similar to the few players that controlled the Nasdaq during the Dot-Com period, the Magnificent Seven control the returns on the S&P 500 index now (accounting for over 30% of the index’s total weight in 2025 (Yahoo Finance, 2026)).

Data from FINE442 Presentation

At the end of the internet-based investment boom, the market collapsed in 2000. Countless companies went bankrupt, investors lost significant wealth, and the NASDAQ fell sharply. With confidence in equities shaken, many investors started looking for safer ground, and with the Fed cutting interest rates aggressively in response, real estate quickly became the most attractive alternative. Which later become one of the core reasons for the 2008 Mortgage Crisis which is another topic that we handled. However, it seems to be diffferent at this time as we can see that the dynamics are quite different from the crisis that we experienced almost two decades ago.

The Dot-Com Bubble post can be found here.

The possible consequences

Yep, you might ask then what happens? We all ask that question and it is pretty difficult to forsee the result, however; we might think of a couple of consequences.

At this moment, there are two distinct possibilities. The first scenario assumes that the AI bubble will burst and have far-reaching effects on the global financial market, while the second suggests that the bubble will never burst at all but simply enter a period of innovation, creative destruction, and gradual repricing.

There are many ways the bubble can burst. Any geopolitical issue, a supply shock or even a stagflationary environment in which increasing costs and shrinking demand are the rule may prompt investors to move away from AI stocks and towards safer assets, such as gold. If tech giants' revenues fail to meet the enormous expectations built into their valuations, or if regulatory pressure increases, the narrative that has been supporting these prices could unravel quickly, cutting off cash flow for startups and potentially triggering a broader downturn.

Even if the AI bubble bursts, the damage would likely be less severe than what we saw in 2000, simply because today's investments are mostly backed by cash and equity, not mountains of debt. Banks would take a hit through bad loans and therefore an increase in their loan loss provisions, but nothing close to the systemic collapse of 2008. The real threat, though, is what happens around it. Moreover, trade wars, supply chain disruptions, rising costs and falling demand can also cause the cuts for the ai investments and additionally, that combination can quickly start looking like the 1970s stagflation all over again. And that puts central banks in an almost impossible position: raise rates to fight inflation, or cut them to support the economy? You cannot do both at once. The other uncomfortable truth is that if they do step in and save the day, they might just be setting up the next bubble.

Conclusion

So, is the AI boom a bubble? Probably not in the classic sense, the underlying technology is real, the leading companies are profitable, and the economic impact is measurable. But that does not mean the market is healthy. There is a growing gap between what AI companies earn today and what investors are paying for them, and that gap leaves very little room for disappointment.

The most likely outcome is probably not a catastrophic collapse but a painful valuation correction, a repricing as the market adjusts expectations to match actual earnings. The companies that survive and thrive will be the ones building real, durable value, not just riding the narrative. The ones that disappear will be the ones that existed only because of the hype. We have seen this story before. The internet did not fail, but most of the companies that promised to define it did. AI will probably follow a similar path.

The question is not whether AI has a future. It clearly does. The question is whether you are holding the right part of the market when the repricing happens.

Thank you for reading and also great thanks to my friends who made this post possible.

See you soon!

This blog has significantly drawn from our article written for FINE 442 (Capital Markets and Institutions) at McGill University's Desautels Faculty of Management, and cannot be regarded as solely personal work but as a team outcome. Co-authors: Mathis Bélanger, Maxime Gauthier, and Tristan Héroux. Full report available:HERE.