The AI Bubble: Why the Champagne on St. Barts May Soon Run Dry

Every time Silicon Valley starts talking about a new era in which the old laws of economics no longer apply, somewhere in the shadows a seasoned investor starts checking his parachute. We are now seeing the classic picture: OpenAI is spending billions on training new models and unicorn startups are multiplying like mushrooms after rain, while skeptics begin asking awkward questions about return on investment. The irony is that technology manias are cyclical.

Look at Britain's "canal mania" of the 1790s and you will see the same symptoms as today. People then believed a network of waterways would connect the world and create endless wealth. The canals got built, but most investors were ruined, and the cream was skimmed by those who arrived on the ruins of their ambitions.

Today's AI industry is not uniform, and it is critically important to understand that. We can loosely divide it into four layers. The first is the "shovel sellers," or chip makers such as Nvidia.

They have already made their billions and are doing very well. The second layer is the cloud giants that rent out computing capacity. The third is the builders of foundation models, such as OpenAI or Google.

And the fourth is applied startups that build their products on top of someone else's APIs. The problem is that most of the money is currently burning at the third and fourth levels, where infrastructure costs often exceed revenue from users. This is the classic growth trap, in which every new customer brings not profit but additional losses.

The situation is complicated by the existence of two parallel realities in AI. The first is the corporate sector, where algorithms quietly and efficiently optimize logistics, find bugs in code and automate tedious bookkeeping. Here the economics work, and companies are willing to pay for real efficiency. The second is "magical" consumer AI: chatbots that compose poems and video generators that demand colossal computing power. This is where the biggest bubble is inflating. Users are accustomed to free or very cheap tools, while the cost of a single complex query remains high. If venture funding dries up before these models become an order of magnitude more efficient, a loud fall awaits us.

Investors in this race can also be divided into several camps. There are the "true believers," convinced that AGI is coming soon and ready to subsidize any losses for the sake of a bright future. There are those driven by pure fear of missing out, who jump into any round with the letters A and I in the name.

And there are the pragmatists, who understand that when OpenAI or another major player faces a liquidity crisis, a great fire sale of assets will begin. That is the moment when capital will be redistributed from dreamers to those who know how to count money. The bankruptcy of a major player will not kill the technology, but it will change the rules of the game for good and force the industry to grow up.

What does this mean for us? Most likely, we will see a series of acquisitions and market consolidation. Small "wrappers" around GPT-4 will disappear because they have no foundation of their own. Those embedded in real business processes will survive, rather than those who merely entertain the public. Technological optimism is wonderful, but it does not replace positive cash flow. When the champagne on St. Barthélemy runs out and investors demand profit and loss reports, we will find out who was really building the future and who was just inflating a pretty balloon. The dot-com era taught us that Amazon and Google were born from the ashes of a burst bubble. It will be the same with artificial intelligence.

The bottom line: should we expect a crash tomorrow? Unlikely. But the era of "free money" for AI startups without a business model is officially coming to an end. Who will be the new Amazon of the neural network world?