Artificial intelligence's massive capital requirements are fueling inflation, according to various economists and Federal Reserve officials. The initial phase of AI infrastructure buildout is particularly capital-intensive, leading to significant demand for resources like specialized computing chips and data centers, which in turn drives up prices. TD Cowen projects major hyperscalers will spend $745 billion this year, with an additional $1 billion-plus in 2027 and 2028, with spending set to reach approximately 3% of GDP next year, a sharp increase from under 0.5% in 2020.
This demand is creating ripples across other sectors; for instance, the scarcity of memory and storage chips used in AI is forcing companies like Apple to raise prices on products such as iPads and MacBooks. New York Fed president John Williams and Fed governor Michael Barr have both highlighted the inflationary impact of AI investment. Minneapolis Fed president Neel Kashkari even shifted his forecast from a rate cut to a rate hike, partly due to the massive data center investments and expected inflation.
The inflationary pressures are expected to continue for the next one to two years during this investment phase. Greg Daco, chief economist for EY, explains that technological revolutions initially raise prices due to intense capital requirements before potentially leading to productivity gains and disinflation in a later phase. This means that the Federal Reserve faces a challenge, as AI's large capital needs are lifting the shorter-run neutral interest rate, potentially necessitating higher rates to maintain price stability, even if AI eventually proves disinflationary in the long term.
