AI has a $2.5 trillion problem and replacing 250 million jobs is the only way to solve it

The rapid development and growing capabilities of AI have led to an unprecedented investment of $2.5 trillion. However, according to new analysis by Marin Ivezic, the founder of Applied Quantum, AI will need to replace 250 million jobs or generate $2.5 trillion in annual revenue in order to justify this investment. Current estimates of AI's revenue are far below this target, with optimistic projections placing AI's current revenue between $150 billion and $250 billion annually.

The $2.5 trillion investment has been divided into $1.5 trillion spent on hardware since 2022 and an additional $1 trillion expected each year. Yet, the US only has 159 million nonfarm jobs, leaving infrastructure providers with significant depreciation costs. High-performance AI chips become obsolete every five years, while the data centers they are installed in are expected to last for 20 years.

This creates a situation known as the '50-Cent Rule,' where providers must generate $0.50 in annual revenue for every dollar they invest. The situation is reminiscent of the dot-com bubble, where infrastructure providers invested heavily in projects that eventually left them holding astronomical costs without a way out. This $2.5 trillion of investment needs to be paid back, and the entire infrastructure market is built on this precarious foundation.