A $3 trillion investment wave is pouring into AI datacenters, splitting experts into two camps: those who see a new industrial revolution, and those who warn of “debt-fueled exuberance” that will “backfire.”
The “boom” camp points to “remarkable numbers.” Nvidia, the AI chipmaker, is a $5tn company. Microsoft and Apple are $4tn companies. Google’s parent just posted a $100bn quarter. In Newport, Wales, a new Microsoft datacenter is creating generational jobs, with local leaders urging the city to “embrace the future.” With 800 million weekly users, ChatGPT’s success seems to confirm the demand.
The “bubble” camp points to the financing. A $1.5tn funding gap for the $3tn spend is being filled by private credit, a “shadow banking” sector. Gil Luria of DA Davidson warns this is financing “speculative assets” and creating “structural risk to the overall global economy.”
This “exuberance” is being questioned by hard data. A recent MIT study found 95% of organizations get zero return from generative AI pilots. Furthermore, the Uptime Institute stated many announced datacenters are “speculative” and “will never be built,” while Alibaba’s chair has warned of a “bubble.”
The industry’s “hyperscalers”—Google, Meta, Amazon, and Microsoft—are spending $750bn, a bet they can afford. But the other $1.5tn in speculative debt, built on “lofty revenue expectations” and “quickly depreciating assets,” is what keeps economists awake at night.
The Great AI Debate: Sustainable Revolution or “Debt-Fueled Exuberance”?
