AI will cost the world tens of trillions of dollars by 2050
Global investment in the construction and modernization of data centers could reach $31.6 trillion by 2050 due to the rapid growth in demand for artificial intelligence technologies. If the pace of AI adoption accelerates, total spending over this period could rise to $50 trillion.
In terms of scale, this boom dwarfs historic infrastructure projects such as railroad development, electrification, or the rollout of the internet. By way of comparison, the annual GDP of the United States is about $30 trillion.
Where Will These Trillion-Dollar Investments Go?
According to the forecast, annual global spending will rise from $800 billion in 2026 to $1.1 trillion in 2030 and $1.8 trillion by mid-century:
Regional breakdown: nearly half of all funds—$15.1 trillion—will be attracted by the United States. The Asia-Pacific region will account for $8.2 trillion, Europe for $5.6 trillion, the Middle East for $1.1 trillion, and Africa for $255 billion. A significant portion of this growth will come from the Chinese and Indian markets;
Expenditure Structure: Unlike traditional infrastructure cycles, the lion’s share of funding will go not toward land or construction, but toward the regular replacement of “hardware” (chips from companies such as Nvidia, servers, and networking equipment), whose replacement cycle repeats every four to six years;
Key risks and challenges: The availability of scalable, carbon-free electricity will be the main factor in facility siting. At the same time, public resistance is growing—in just the first three months of the year, protests blocked or delayed 75 projects worth $130 billion. In addition, disruptions in semiconductor supply chains could reduce the projected investment volume by 20%.
This is reported with reference to data from Bloomberg.
China has deployed an autonomous AI system at drilling sites.
Genesis BBQ Group plans global expansion using AI — Korea Herald.
A U.S. House of Representatives subcommittee has advanced a bill on open AI models.
