Michael Dell, founder of Dell Computers, said recently on LinkedIn:

“The skepticism around AI’s ROI today feels a lot like the doubts people had about the internet in its early days. Back then, no one could fully grasp how much the web would transform our lives, and now AI is on the same path. Yes, it might be tough to measure immediate returns, but if history teaches us anything, it’s that game-changing technologies take time to show their true impact. By embracing AI, we’re not just adopting a tool — we’re investing in a future where human potential gets a major boost, and new opportunities are unlocked. Just like the internet became the foundation of our digital world, AI has the potential to do the same, when we approach it with the same optimism and vision.”

Hmmm… Michael Dell must be suffering from amnesia or even Alzheimer’s, because the early days of the Internet were very different from today’s early days of AI. There was an exploding market for Internet services, from e-commerce to content, during the dotcom bubble, while today the market for AI services is hard to find. This is why Sequoia’s David Cahn, Goldman Sachs’ Jim Covello, Citadel’s Ken Griffen, and long-time bull Ed Yardeni have talked many times about the AI bubble.

Looking Back at E-Commerce

A March 2001 article claims the market for e-commerce was $286 billion in 2000, which in 2024 dollars is about $500 billion. Cahn, Covello, Yardeni, and Griffen are complaining because the market for generative AI services are in the few billions when they need to be in the $500 billion dollar range, $600 billion according to Cahn, to justify the expenditures on Nvidia chips. 

That $286 billion figure included many items that we continue to buy; the top two in 2000 were travel and computers, followed by flowers. Many of us can still remember booking hotels and airline flights two decades ago and I can remember buying flowers for my mother. Even more people make these purchases and bookings today, along with buying computers, books, food, music, clothing, and thousands of other things online. 

Ask yourself, how many of today’s AI applications can even remotely compare to the convenience we got from buying things on the Internet in 2000? Could you imagine being forced to go through a travel agent for all your trips today? Maybe if you have a dedicated assistant but not if you are traveler who wants to get a decent deal. 

Even the first items purchased with e-commerce, books, are as good an application or better than most of today’s AI applications such as writing e-mail or reports or generating ideas. I remember the challenge of finding specialized books before the Internet and the joy of finding them on Amazon. Another memory for academics: remember the library copying services of the 1990s that enabled us to read hard copy journal articles in our offices? 

The Dotcom Bubble was Mostly Big Incumbents, not Startups

A second mistake is to assume the dotcom bubble was mostly about startups with little or no revenues such as Pets.com and Webvan, when in reality the Nasdaq bubble mostly consisted of big companies such as Cisco, Qualcomm, Nokia, Dell Computer, and Intel, most of whose market capitalizations have not returned to the levels of 2000, or at least they didn’t until the late 2010s. There was also too much invested in telecom infrastructure during the Internet bubble just as there has probably been too much invested in today’s big cloud providers, Google, Microsoft, and Amazon, which is the concern of Cahn, Covello, Yardeni, Griffen, and others.  

We forget these big incumbents because we mostly remember Amazon, Google, Nvidia, and other startups that have since achieved huge market capitalizations. These startups represented very little of the dotcom bubble just as new startups represent very little of the current AI bubble, other than a few privately held startups such as OpenAI and Anthropic. That in itself should be a concern, because it is possible that few if any important startups will emerge from this AI bubble. 

The AI Bubble Started More than 10 Years Ago

The AI hype started long before generative AI appeared on the scene. Books such as the Race Against the Machine, the Machine that Changed the World, World Without Work, and The Rise of the Robots started appearing more than 10 years ago. IBM Watson was going to remake healthcare, because, well, it could beat the best Chess and Jeopardy players, just like the latest ChatGPT version, i.e., Strawberry, can effectively recite economic theory according to Tyler Cowen. And projections of a $16 trillion market for AI by 2030 had appeared by the mid-2010s. 

When you hear that AI is in its early days, be skeptical. Proponents of AI have been saying this for years if not decades. We are now on the third wave of AI hype (the second was while I was a graduate student at Carnegie-Mellon in the early 1980s) and the third wave started more than 10 years ago.

Ironically, the hype for AI had almost died in 2022 following the bursting of the Unicorn startup bubble. Then generative AI was trotted out and the real bubble arrived. The market capitalization of the Magnificent Seven (Apple, Microsoft, Alphabet, Amazon, Nvidia, Meta, and Tesla) rose by $10 trillion between January 2023 and their peaks in 2024. Although there have been some subsequent declines, the share prices for the Magnificent Seven are still near their highs.

There are a lot of ways to think of AI. You can compare the sizes of AI software to e-commerce, the investments in infrastructure, the rises in the market capitalization of the big tech companies, or the impact on our lives of the Internet and AI. In the important ways, today’s AI is a poor performer compared to the Internet. 

Michael Dell should know better, as should other tech bros. They should be talking about the present, and not just the future. Because long-term forecasts should be consistent with short-term trends.

This is the subject of my book Unicorns, Hype and Bubbles, forthcoming on October 22.