The next wave of the boom is based on debt, and comparisons with the dot-compor bula are multiplying

Foto del autor

By TP


Openai prepares the land for a network of data centers that will cost at least $ 1,000 billion for the coming years. In the first years of the boom, the comparisons with Bula dot-Com did not have much logic. Now, after three years, the growing levels of debt make such comparisons sound a little more true, writes The Wall Street Journal. At first, the rich Tech companies opened their wallets to compete in winning the leader in AI. They spent cash generated largely from advertising and cloud computing divisions. There were no massive expenses based on infrastructure debt like those that swelled the bubble two and a half decades ago. While the big Tech companies are still in the first lines of AI and are financially shaped, a number of companies with a higher level of debt open the way to an era that could change the nature of the boom. Some smaller companies, led by Coreweave, are based on a more creative financing to propel the first plane for a period. More recently, however, the ambitions of some companies like Openai are about to carry the indebtedness and mega contracts at a completely new level. Openai prepares the land for a network of data centers that will cost at least $ 1,000 billion for the coming years. As part of the initiative, the company has signed a $ 300 billion contract, for five years, this month in which Oracle will create a calculation infrastructure based on AI and rent it to Openai. In order to honor its contract, Oracle will have to spend on that infrastructure before being paid entirely by the openii, which means many loans. Keybanc Capital Markets analysts estimate that Oracle will have to borrow $ 25 billion a year in the next four years. And Oracle already has a high degree of indebtedness. Oracle and other less wealthy aspirants at have few other options than borrowing if they want to play in large leagues. For now, investors are not too worried. However, there are reasons for skepticism related to the ability of these companies to fulfill their contracts and pay their debts. Numerous recent studies have found that you do not gain land as quickly as its supporters suggest: only 3% of consumers pay for it, a study reveals. The estimates that the expenses on data centers will reach thousands of billion dollars a year in the coming years seem extremely optimistic. There is a decent chance for everything to go well. Oracle could handle their contracts and debts well, and players like Coreweave could bring a wave of financial innovation that will boost the development of AI. But it is at least as likely that many of the massive contracts today be postponed have renegotiated because the final demand for services will not increase in step with the development of infrastructure. The contracts could be transferred, lawyers say. If Openai has problems, Oracle could rent the infrastructure of a more stable company, assuming the agreement allows this. Such a situation would not be a death blow to Oracle and the other indebted companies, but they would be a formidable test for a financial model at the beginning, one that looks more and more like a bubble on the passing day.
For other news, analyzes, articles and information from real -time business follows the financial newspaper on WhatsApp Channels