How can a technology be revolutionary and, at the same time, generate a financial bubble around it? What seems to be a contradiction is the distinction that Scott Galloway , writer, entrepreneur, and professor at New York University, makes in his most recent text.

Considered one of Silicon Valley's gurus, he has spent the last few years among the most optimistic voices about the impact of artificial intelligence. He has repeatedly stated that AI represents a paradigm shift comparable to the internet.

Now, while Galloway remains a proponent of AI as a transformative technology, he has come to make a very clear distinction between the technology's potential and the price of the assets associated with it.

In the text "1999.AI", Galloway states that signs of the dot-com bubble are reappearing. Not because artificial intelligence is a fraud or because the technology has failed.

On the contrary: a revolutionary technology can perfectly coexist with a financial bubble. "I remain as optimistic about AI as I was about the internet in 1999," he writes. "But I've learned, with some scars, not to confuse valuation with value."

This is perhaps the most important distinction in the current artificial intelligence cycle. There is a difference between believing that AI will transform the economy and believing that all companies linked to AI justify the ratings they have received.

That was the case during the dot-com bubble, which burst in the second half of the 1990s. Investors began to believe that any company with ".com" in its name was destined for success.

In 1999, nearly 40% of venture capital investments in the United States were concentrated in internet companies, while four out of five IPOs had some connection to the new economy. Capital stopped distinguishing between good business models and good stories.

The internet has changed the world, although most of the companies from that first generation have disappeared.

Galloway cites the example of Pets.com. The thesis was that consumers would indeed start buying products for their pets online. Right idea, wrong timing.

The infrastructure was still precarious, costs were high, and the financial model simply didn't work. A decade later, the same idea would resurface in companies like Chewy, this time in a much more favorable environment.

According to Galloway, artificial intelligence may follow a similar script. In other words, his criticism is not directed at the technology itself, but at the excessive expectations embedded in the valuations of the companies leading this race.

He cites OpenAI's billion-dollar losses, the high cost of operating the models, and the gap between revenue projections and actual results.

Moreover, it draws attention to an aspect that it considers typical of the final stages of major bubbles: the growing dependence on a continuous flow of capital to sustain a model that is not yet generating returns commensurate with the investments made.

During the dot-com bubble, the first victims were consumer-facing companies. Then came software and advertising providers. Finally, the telecommunications infrastructure giants, which had invested billions to meet a demand that quickly disappeared.

Similar chain

Currently, the scenario is no different. Companies invest billions in artificial intelligence to avoid falling behind. This spending fuels the accelerated growth of model developers, who in turn justify equally massive investments in chips, data centers, and computing infrastructure.

But this mechanism depends on increasing corporate spending. According to Galloway, the first signs of change have already begun to appear. Large companies have started to limit token consumption, review AI projects, and demand measurable financial returns before expanding investments.

The phase of "using AI because everyone else is using it" is beginning to give way to demanding effective productivity. This has, in recent months, united a growing group of economists and investors, all of whom have come to defend a similar thesis.

Aswath Damodaran, a professor at NYU known as one of the world's leading authorities on valuation , argues that the risk lies not in the technology itself, but in the extraordinary expectations embedded in AI-related stocks.

Jim Covello, an analyst at Goldman Sachs, questioned whether the hundreds of billions of dollars invested in infrastructure will produce sufficient economic return to justify this level of spending.

Nobel laureate in Economics Daron Acemoglu has also adopted a cautious stance. He believes that artificial intelligence will produce significant productivity gains, but at a much slower pace than the market currently prices in.

Perhaps the closest warning to Galloway's came precisely from Silicon Valley. David Cahn, of Sequoia Capital, summarized this concern in the article " AI's $600 Billion Question ": who will generate enough revenue to pay for all the infrastructure built to support this revolution?

Galloway has an answer (perhaps a somewhat provocative one and far from what the financial market expects). He compares artificial intelligence to electricity, a technology that has revolutionized virtually every sector of the economy, but whose greatest benefit has ended up being distributed among millions of consumers and businesses, and not concentrated in the technology manufacturers.

On the internet, Amazon, Google, and a few other winners became gigantic. But a large part of the value created was captured by traditional companies that learned to use the network to sell more, produce better, and reduce costs.

It is entirely possible that AI represents the greatest technological revolution since the internet — and, at the same time, that a significant portion of the companies currently considered winners will discover that the market paid too high a price for a future that will take much longer to arrive.

The internet survived the Nasdaq crash in 2000. Galloway wants to know who will still be standing when the dust settles—and the bubble bursts.