Analysis:AI's race to transform the world before the money runs out

AI technology is experiencing an unprecedented surge in investment, dwarfing previous technological revolutions such as railways and the internet. Data centers alone could account for $30 trillion in global spending by 2050, surpassing the capital invested during railroad and dotcom booms. Major players like Anthropic plan to invest $518 billion in the coming years, far exceeding their 2025 revenue. However, the assumptions behind vast productivity gains and future profits remain speculative and lack historical precedent.

Economists question whether AI can deliver the promised productivity gains, with JP Morgan noting that broad-based productivity gains in the US, the leader in AI, remain elusive. Bain & Company suggests that entirely new markets must emerge to justify current AI outlays, such as using AI-guided robots or developing new materials for batteries and semiconductors.

US hyperscalers, including Google, Amazon, and Microsoft, need to generate over $4.2 trillion in new revenue over the next five years to fund AI infrastructure buildouts.

While the potential of AI to transform various aspects of life is widely acknowledged, economists are working out the implications for the global economy beyond investment cycles. Historical precedents suggest that technology-driven booms often end when infrastructure buildouts fail to deliver sufficient returns. For the US, AI investment could reach $9 trillion from 2025 to 2032, equivalent to spending 3.2% of US GDP annually.

To justify its valuation, the US AI sector would need a 3% to 5% annual productivity gain over the next decade, significantly higher than the Congressional Budget Office's baseline expectation of 1.75% annual growth.

The potential for recursive self-improvement, or AI systems learning to improve themselves, is seen as key to achieving exponential advances and unprecedented productivity gains. However, concerns about existential risks to humanity linger. Despite the dizzying projections, economists argue that the pace of productivity change may still lag behind corporate timelines.