AI’s trillion-dollar bet: Will productivity gains arrive in time?

Since the onset of the AI revolution, capital investment in this technology has surged like never before, surpassing previous technological revolutions such as railways and the internet. According to a projection by PwC, global spending on data centres could reach $30 trillion by 2050, nearly matching the outstanding value of US Treasuries.

Even the spending of major firms like Anthropic, which plans to invest $518 billion in the coming years, dwarfs its 2025 revenue by more than 100 times. While proponents of AI technology tout its potential to revolutionize everything from office work to research labs, similar to past technological advancements, economists caution that the assumptions behind projected productivity gains and future profits lack historical precedent and evidence.

JP Morgan highlights the elusive nature of broad-based productivity gains in the US, which leads the AI race, questioning the sustainability of AI valuations. To fund the infrastructure buildout, US hyperscalers and other players in the AI race may need to generate over $4.2 trillion in new revenue within the next five years, as per a Bain & Company study.

While the potential of AI to transform various aspects of life is undeniable, economists emphasize the importance of securing a return on investment and meeting deadlines for repaying loans, which ultimately influence the global economy beyond investment cycles.