AI’s $30 Billion Bet Sparks Inflation and Debt Fears as Data Center Demand Outpaces Supply
Economists and investors are warning that the rapid adoption of artificial intelligence (AI) could drive up prices and strain financial markets if data-center construction cannot keep pace with demand for computing power.
The AI boom is putting unprecedented pressure on the infrastructure that supports it. SoftBank is planning to borrow $30 billion to fund large-scale AI infrastructure, including data centers and custom chips, marking one of the largest financing efforts in the sector to date [267045]. However, the Japanese investment giant already carries a heavy debt load from past investments, and analysts say the success of this strategy depends on whether SoftBank can generate steady returns before the debt becomes a burden [267045].
The warning extends beyond individual companies. Alicia Garcia Herrero, chief Asia-Pacific economist at Natixis, said that a slow buildout of AI infrastructure could create inflationary pressure [271055]. If demand for AI computing outpaces the supply of data centers, costs could rise across the economy [271055]. Her warning suggests that the speed of AI adoption must be matched by investment in the infrastructure that supports it [271055].
Investors are also starting to see the risks in debt linked to AI. Traders are now pointing to the weak spots in AI-related bonds, showing where the biggest dangers may be [266331]. Meanwhile, Nvidia has evolved from a chipmaker into a major financier of the AI industry, and some investors worry the company is taking on excessive risk tied to the AI buildout [270986]. The recent market decline of neoclouds—smaller cloud providers that rent AI computing power—is testing this new role for the first time [270986].
The infrastructure strain is not limited to finance. Malaysia is racing to become a top destination for data centers, but this rapid growth may push the country's electricity supply to its limit [270302]. Data centers use huge amounts of power, and as more centers open, demand for electricity will rise sharply [270302]. Without careful planning, the data center boom could lead to power shortages [270302].
The pressure extends to the internet's underlying systems as well. Experts say the internet is held together by makeshift fixes and fragile systems, and AI tools demanding massive computing power and data could expose weaknesses in the networks people rely on every day [270750]. The result may force a long-overdue rebuild of the internet's core [270750].
---