$3Tn AI Spend: “Lofty Revenue Expectations” vs. “Zero Return” Reality

The $3 trillion AI datacenter boom is running on “lofty revenue expectations,” with Morgan Stanley projecting the generative AI market to hit $1 trillion by 2028. But a “rattling” MIT study shows a “zero return” reality for 95% of businesses, creating a massive disconnect that could signal a “bubble.”
These “lofty expectations” are what justify the $3tn spend. They are why “hyperscalers” (Google, Microsoft, etc.) are spending $750bn in two years. They are why Nvidia is a $5tn company and OpenAI is valued at $500bn.
But the MIT study shows that business “takeup” is failing. If 95% of organizations are getting “zero return,” they will not become the paying customers needed to reach that $1tn market projection. This is the core problem.
This “exuberance” is funding a $1.5tn “speculative” boom, financed by “private credit.” These lenders are funding “unproven” projects “without their own customers,” based on the same “lofty” expectations. Alibaba’s chair, Joe Tsai, has explicitly warned of a “bubble” in these projects.
The industry is betting $3 trillion that the “zero return” figure is temporary and that the 800 million weekly users of ChatGPT will translate into a $1tn business market. If they are wrong, the “speculative” debt-fueled boom “will backfire.”