AI skepticism reaches Asia: investors no longer want to pay for promises

The seemingly endless line of credit the stock market extended to artificial intelligence has started to run out. What began as modest profit-taking in Silicon Valley has turned into a full-blown flight from tech assets, which has now reached Asian exchanges too. Investors who only yesterday were ready to buy any stock with "AI" in its description have suddenly sobered up and started counting their money. The question of the day sounds frighteningly simple: where is the promised revenue from all those billions invested in computing power?

The situation resembles a classic market hangover after a long party. For the past year and a half, the industry lived in gold-rush mode, where the main asset was the shovels, meaning chips from NVIDIA and its rivals. Now, however, the focus is shifting from those who build the infrastructure to those who are supposed to earn money on top of it. When Microsoft, Google and Meta* spend tens of billions of dollars on data centers, shareholders want to see more than polished demo videos on Twitter. They want concrete figures in profit and loss statements. So far those figures are not impressive, which triggered the current wave of skepticism.

Asian markets reacted especially sharply because the key manufacturing nodes of the AI economy are concentrated here. Shares of component suppliers and equipment assemblers began losing value, following their American customers. Investors realized that if Big Tech in the US decides to slow hardware purchases because software products pay back poorly, everyone along the chain will suffer, from makers of lithography machines to chip-testing plants. This creates a domino effect that is hard to stop with simple statements about a bright future.

The problem also lies in so-called "bubbly" valuations. When company multiples soar on expectations alone, any stumble in growth is perceived as a catastrophe. We are now watching exactly this process of repricing. The market is trying to find the bottom and work out how much of company value rests on real business and how much on pure hype around neural networks. It is important to remember that the technology itself is not going anywhere, but the financial landscape around it is clearly changing.

Many analysts draw parallels with the dot-com crash, but the situation is somewhat different. In 2000, companies often did not even have a business model, whereas today's tech giants have huge cash flows from their core businesses. Even so, pressure on AI divisions will only grow. Companies can no longer justify enormous capital expenditure simply by the need to "stay in the race." Now they will have to prove that their products built on large language models are truly needed by customers, and that customers are willing to pay for them.

In the coming months we will most likely see the market split between those who managed to monetize AI and those who simply burned investors' money. This is a healthy cleansing process, although for many portfolios it will be quite painful. The era of easy money is over, and the AI industry now has to grow up and learn to live under strict financial discipline, where every dollar spent on GPUs must come back with a profit.

Bottom line: The market has stopped believing in magic and started demanding Excel spreadsheets. Will AI developers be able to show real profit before investors finally lose patience?

*Meta is designated an extremist organization and banned in Russia.