TThese fears have been reinforced by bad news about the initial public offerings (IPOs) of several high-profile AI companies, when a private company begins selling its shares to the general public for the first time.
First, OpenAI, the company behind ChatGPT, delayed its IPO due to internal concerns that it might end in failure, because retail investors — basically, you and me if we go public — might not be as eager to buy its shares as they hoped. The reasons include general market uncertainty caused by the war in Iran, growing skepticism of AI, a series of damaging headlines about the technology’s security concerns, and a stagnant user base for ChatGPT. The company had to reduce its valuation for its latest funding round due to a palpable decline in investor enthusiasm.
Meanwhile, Claude’s creator, Anthropic, has headlined a prospectus – the document published by a publicly traded company that explains the company and lays out its financial reality to potential investors – that appears to hold true for all of the concerns just exposed about the AI business model. The latest version of the document, obtained by the The Wall Street Journal This week, the company reported a loss of more than $80 billion this year, as its spending on IT and infrastructure increases exponentially, its demand for computing capacity exceeds that available, and its list of possible risk factors is longer than the company’s description.
One of these risk factors has become a major general concern for AI investors: the public’s growing hatred of data centers and the desire to restrict, or even stop, their construction altogether. Popular distaste for data centers cuts across party and ideological lines and has become a central issue in the upcoming midterm elections, making an intensification of government repression against them more likely.
In the first half of this year alone, environmental activists and local residents blocked more than a hundred data centers worth approximately $198 billion in capital. This included a major Oracle project in New Mexico that prompted the company to declare force majeure – that is, tearing up a contract due to force majeure – after having already spent millions.
This is yet another, and particularly important, reason why AI investors are worried about getting their money back. For AI companies to have any hope of meeting their already unrealistic revenue goals, data centers to have to be built somewhere; but voters across the country are fiercely opposed to their integration. their neighborhoods, because they find them noisy, intrusive, harmful to the environment and have the habit of driving up their electricity bills. They therefore organized themselves to restrict them, or even outright ban them, which they have done in more than five hundred counties and municipalities. Something will have to give here.
Not to mention investors’ fears of being left with the burden of data center debt. The Brookings study mentioned earlier estimates that building an AI data center will cost a total of $10.3 trillion by 2032; Market research firm Gartner estimates that AI data center owners will need to generate a staggering $8.5 trillion in revenue by 2030 to cover construction and still earn a decent return. These are huge sums of money behind something that is largely financed by massive and rapidly growing levels of borrowing.
Added to this is the fact that the chips that AI companies rely on and spend billions of dollars on are becoming obsolete more and more quickly – much faster than the lifespan many of these companies initially expected. This is another case of wishful thinking on the part of the industry, and another reason to worry investors, since AI companies’ loans are often secured on the back of chips that will lose their value before the end of the loan’s life.
No wonder investors are now turning to AI and hedging their bets in anticipation of a bubble bursting. It could happen sooner than we think: Oracle, seen as a “bellwether” for the broader AI credit market, has seen its bonds downgraded to just above junk status. One economist believes that if its bonds reached this nadir, it would be a “sign of contagion” for the broader sector.
Gn bussni
