Economic analysis of artificial intelligence bubble; Is another historic recession coming?

Experts say that the artificial intelligence bubble is not an ordinary bubble, but perhaps the biggest one we have ever seen; Something like the ideal, platonic version of a tech bubble that can swallow all bubbles. In this article, based on an economic framework, we take a close look at the artificial intelligence bubble; Will it explode one day and we will face an economic recession?

Ever since the viral success of ChatGPT in late 2022, when every company in Silicon Valley tilts its way to AI, it feels like a bubble is slowly deflating. In recent months, financial analysts, independent research firms, and even the executives of some AI companies themselves agree that we are dealing with some kind of bubble.

But is this assessment accurate? To find a reasoned answer, one should go to the researchers who wrote the reference book on technology bubbles.

Brent Goldfarb and David A. Kirsch, University of Maryland economists, in their book “Bubbles and Crashes” present a four-factor framework for evaluating technological innovation bubbles. Their framework is based on the principle that the narrative and story that develops around a technology plays a central role in the creation of the bubble.

According to this framework, four factors must be present to create a perfect bubble: Uncertainty, Exclusive actors, Novice investors and Coordinating narratives. By applying this test, artificial intelligence gets the highest score in all four cases.

In the following, we will discuss these four factors in the field of artificial intelligence.

1. Uncertainty

Uncertainty is the cornerstone of a tech bubble. Uncertainty about what the technology can do, what the business model is, and how long it will take to come to fruition.

Alarm bells are already ringing about artificial intelligence. The ultimate goals are very vague; So far, no one knows what the ultimate goal of artificial general intelligence (AGI) or superintelligence is and what exactly it is going to do with our world.

Nearly three years on, the major players (except for Nvidia) have yet to prove their long-term business models. Companies are burning billions of dollars, their costs haven’t come down, and most companies are losing money with every user query of AI. A recent MIT study also showed that 95% of companies that have used productive artificial intelligence have not benefited from it.

Goldfarb says uncertainty usually diminishes over time, but that hasn’t been the case with artificial intelligence. Early claims of its effectiveness have been, at best, a mixture of fact and exaggeration.

The closest historical analogue of artificial intelligence is radio. When RCA started broadcasting, it was clear that the technology was powerful, but the business model was unclear. As a result, radio became one of the biggest bubbles in history, losing 97% of its value in the crash of 1929. RCA at that time was the Nvidia of its time.

2. Exclusive actors

Creating a bubble requires companies whose fate is completely tied to the success of that particular innovation. These are exclusive actors. For example, Tesla is an exclusive player in electric vehicles, while Toyota is not; For this reason, Tesla’s valuation far exceeds its earnings.

In the field of artificial intelligence, according to the Silicon Valley Bank report, 58% of the total venture investment this year has been allocated to artificial intelligence companies. At the top of the list is Nvidia, which has tied its future to making chips for artificial intelligence. Companies such as Anthropic and OpenAI are also exclusive players.

What’s troubling about AI is the tight entanglement of these players: Nvidia invests in OpenAI, while OpenAI depends on Nvidia’s chips and Microsoft’s computing power to survive.

3. Novice investors

At the height of the dot-com bubble 25 years ago, retail investors poured their money into companies like Pets.com. Today, hordes of novice investors are pumping money into AI through apps like Robin Hood. For example, in 2024, Nvidia was alone at the top of stocks bought by retail traders.

Brent Goldfarb points out that when it comes to AI, everyone is kind of a novice investor because the technology is so new that no one really knows how it will pan out. Economists say investors have sunk their savings into the vague promise of superintelligence.

4. Coordinating narratives

This is the strongest factor in the AI ​​bubble. In 1927, Charles Lindbergh’s solo flight over the Atlantic created a powerful narrative that signaled investors to pour their money into the aviation industry. The result was a bubble that burst in 1929 and aviation stocks fell 96 percent.

In the case of artificial intelligence, this narrative is oddly more powerful. Industry leaders are promoting a narrative that AI will soon do everything humans do: jobs will be automated, cancer will be cured, climate change will end, and AI will literally be “everything.” Adding the “we must defeat China” narrative also adds more fuel to this fire. Goldfarb says, “Does this narrative work? The answer is a resounding yes.

Brent Goldfarb confirms that AI has all the hallmarks of a classic bubble: “Without a doubt, it ticks all the boxes.” Uncertainty? yes Exclusive cast? yes Newbie investors? yes Great narrative? yes He says on a scale of 0 to 8 of this framework, artificial intelligence Full score 8 receives

The phenomenon’s closest historical analogs (radio and aviation) both helped burst a bubble so large that it led to the “Great Depression” of 1929. History repeats itself, and this time, “the buyer must be aware.”