AI summary
Summary of article:
Kluman believes that although US AI investment seems unprecedented, it will cool down due to financial difficulties and public reaction to prevent the bubble from getting out of control, but some companies may still face bankruptcy and investment losses.
According to AI-related spending plans announced by major players in the artificial intelligence (AI) industry,columbia university Economist Van Nieuwerbergh predicts this will be the largest investment boom in US history, measured by the ratio of AI capital spending to gross domestic product (GDP). is that true?
2008Nobel Economics Award Winnerkrugman Bold prediction on the 29th: It won’t happen. On the contrary, this AI investment boom will hit two walls, and most projects will either be postponed or canceled.
Kluman pointed out that the first wall is the lack of funding needed for AI investment. So far, most AI spending comes from technology giants like Google and Microsoft. These technology giants initially relied primarily on profits generated from their existing businesses to finance AI spending; But now the scale of AI manufacturing has become so large that they need to rely on large amounts of debt to support it.
Lenders are slowing down. Many observers claim that AI is a great technology, but its profitability prospects are unclear. It now appears that the millions of dollars that companies have spent on cutting-edge AI technology are unlikely to generate enough profits to justify such large expenditures. There are many reasons. The most obvious competition is from mainland China’s AI models. Mainland China’s AI products are cheap and have weak functions but are good enough for commercial use.
In fact, this situation sounds familiar. The boom in information technology (IT) investment in the late 1990s changed the world, but it also led to the bankruptcy of many major technology companies because they could not find a way to monetize their technology. Over the past few weeks, it seems even the top AI companies have suffered a “Willy Wolf moment” – falling off a cliff and crashing to the ground.
Financial data also shows that skepticism is spreading in debt markets. Credit default swap (CDS) spreads – the price of insurance bought to protect companies from debt – at many AI companies began rising this summer. CoreWeave, whose core business is selling computing power, is at particularly high risk of debt default, but other companies are also beginning to come under pressure, especially Oracle.
Scared lenders have two reactions: demand higher interest rates, or simply refuse to make the loan. Under such circumstances, it is becoming difficult to imagine how hyperscalers can realize their huge investment plans unless one is willing to bear huge losses.
The boom in AI investment also faces another obstacle: Most Americans have a deep hatred of data centers and AI in general.
Surveys show that people across the United States are very frustrated with data centers. Not only do most Americans oppose data centers being built near their homes, but most respondents also oppose data centers being built elsewhere in the country. Moreover, the opponents far outnumber the supporters.
There are many reasons for this hostility. Data centers consume large amounts of water and electricity, which not only increases the cost burden on the community, but can also cause pollution when building self-built power supply facilities. Also, once the data center opens, it will bring almost no job opportunities to the local area.
Furthermore, people are generally worried that AI will lead to mass unemployment and even the extinction of the human race – this kind of doubt comes from the mouths of AI industry executives, who believe that the spread of “doomsday theories” is conducive to their initial public offering (IPO).
Typically, technology giants compete to build data centers, usually through “backroom deals” with corrupt politicians. This deepens the public’s dislike of the rich and powerful technology elite, who they believe benefit themselves at the expense of ordinary people.
Now, American public opinion is coming back in a big way. Grassroots protests have had a major impact because building huge data centers requires permits, which more and more local governments are refusing to grant due to growing public outrage. The main cause of Oracle’s problems is the “Jupiter Project” to build a super-large AI data center park in New Mexico. The natural gas pipeline permit for power supply has not yet been approved.
