Written By : Vijay Jog ( Founder, CRG Solutions)
What defines a bubble is subjective, and when it would burst (but burst it does) is unknown, especially when one is in the bubble. A  bubble arises because of “irrational exuberance” where prices of a stock or a commodity or a company or an idea detaches from “fundamental” value. We have had at least 10 such bubbles to date – beginning with the now famous Dutch Tulip mania bubble in the 1630’s to the one that some of us above 40 years of age must remember  – the infamous Dot.Com bubble (late 1990s – early 2000s ).  This bubble was global in nature but impacted the US financial market participants the most.  The bubble lasted longer because of two behavioural characteristics displayed by mortal human beings: FOMO and the Greater Fool Theory (GFT) meaning “there will always be another fool willing to buy from me at a higher price.”

During the Dot.Com bubble, we saw rapid investment by wealthy angel investors, venture capitalists, and even pension funds in companies with very unrealistic business plans and projections leading to inflated valuations where the name of the game was to burn money fast, compete for a share of “eye balls” and have a site/company with a .com domain.  During  this period, investors deviated from the age old and proven metrics such as price per revenue, price to earnings or EV to EBITDA because these companies neither had revenues nor earnings – all they had was continued burn rates and highly paid 20-year-olds who claimed that they would change the world and everyone would become a billionaire. Some companies even went IPO with underpricing of 200% to 300% on the first day of trading. The bubble officially peaked around March 10, 2000 when the Nasdaq Composite index, heavily weighted with technology stocks, hit its all-time high of 5,048.62 and reached a low of 1,108.49, a 78% decline from peak to trough, and it then took fifteen years to reach the peak, which it did finally on April 24, 2015.

The bubble “burst”  for a variety of reasons including macro (increase in the US Treasury yield), but also because of micro reasons – unrealistic business models and revenue/profit projections demonstrated by mediocre performance, including those by  B2C retail companies in the Christmas of 1999 and 2000 (example being: Pets.com, Webvan, Boo.com, eToys.com: Flooz.com) etc.  In some cases, consumers were not ready because today, we do have companies with similar business models and they are now profitable. The Dot.Com bubble burst destroyed market capitalisation of even the most established companies such as Amazon (AMZN): from a high of around $107 to a low of $7 (90% decline), Cisco Systems (CSCO): (over 80% of its stock value), Intel (INTC): Its stock price dropped almost 84%, and even Microsoft (MSFT): about 60% decline. However, the lessons learned from the bubble and the leveraging of the internet infrastructure led to successful companies such as Zillow, Amazon, eBay, and Priceline.

So not all was lost.

I am now seeing some parallels in claim about this AI revolution. It is a new paradigm that the world has never seen before and that we have learned enough from history to avoid another bubble again. However, one should also note that we have had a few “bubbles” since the dot-com bubble: The U.S. Housing Bubble (United States, 2000s, and bursting around 2007-2008), the Commodities Bubble (Global, 2007-2008) and the Chinese Stock Market Bubbles (China, notably 2007 and 2015) and Cryptocurrency Bubbles (Global, various periods, notably 2017 and 2021) where the same sentiments were expressed.  Also, there are some eerie statistics that I can cite to show that we may be in an .AI bubble now.

Let me start with  the .AI domain explosion. Let us focus on the Internet top level (TLd) domains, but ignore the country or industry or location specific, second or third level and Infrastructure or test top-level domains.  These Top-Level Domains include   .com, .org, .bz,.net, .org  and now the new kid on the block, that is  .ai.

According to Verisign’s “the Domain Industry Brief Quarterly report Q1, 2025”, there are 169.8 million of .coms (157. Million) and .nets (12.6 million) LTD domains. Naturally, The United States has the majority of these domain registrations (over 75%) because it originated there. India has around 1.8 million .com domains representing about 25% of  all registered domains in India, with the .in ccTLD (country code Top-Level Domain) being more prevalent. The popular choice in India is co.inand is also available to anyone worldwide and is often seen as a more business friendly option compared to just .in.

The first dot.com domain, symbolics.com, was registered on March 15, 1985. Only six additional .com domains were registered in that year. By 1995, the numbers surpassed 50,000 and by year 2000, it increased to 1.5 million and by 2010 it was 100 million, and 145.5 million by 2020. The .com domain has consistently remained the most popular and trusted domain extension.

Now let us compare that with .AI domain registrations. While the .ai domain technically originated as the country code top-level domain (ccTLD) for Anguilla in 1995 and it continues to derive revenues from its use, this significant growth is directly tied to the artificial intelligence fever of late. The use of .AI really started in 2018 with  approximately 50,000 .ai domains being registered, and since then, it has exploded with year over year growth reaching 1005 +. As of early 2025 (Q1), there were 610,000 .ai domains and the number grows each month by 20,000. The popularity has grown to a level that Google now treats .ai as a generic top-level domain (gTLD). This means that it’s not geographically restricted in search rankings, which further boosts its appeal for global AI-focused businesses. India accounts for approximately  80,000   domains (approximately 14% of the total). Note that this number represents domain registrations, not necessarily unique companies. A single company might own multiple .ai domains.

Next, let us look at market cap of companies that claim to be “all in” in AI.  The Magnificent Seven have market capitalization of close to 15 billion.  Then next tier has companies like Palantir Technologies (PLT), C3.ai (AI), SoundHound AI (SOUN), Innodata (INOD), Quantum Computing Inc. (QUBT), and the statistics are as follows:

And of course, we have other private companies with valuations ranging between $2-6   billion (Perplexity AI, Synthesia, Mercor, Harvey, Cohere) to $60 billion (Anthropic) along with a large influx of venture capitals, which some estimate would reach $100 billion by year end, including a sizable portion to companies in the GenAI space. High valuations in acquisitions are also eerily similar – examples being  Microsoft’s non exclusive acquisition of Inflection for USD 650 million and Alphabet’s USD2.6 billion acquisition of Windsurf.

And I am not alone. This is how Grok sees it