The rapid expansion of AI-driven data centers is emerging as a significant new source of inflation, primarily by increasing demand for electricity and other core resources. The sheer energy needs of these facilities, with some requiring as much power as a major city, are driving up retail and wholesale electricity prices. For instance, the Federal Reserve Bank of Dallas projects that under plausible assumptions, annual Personal Consumption Expenditures (PCE) inflation in the U.S. could rise by an additional 0.04 to 0.13 percentage points in 2030 due to data center electricity consumption.
Beyond electricity, data centers are consuming vast amounts of land, water, labor, and building materials, creating scarcity and driving up costs across multiple sectors. The construction of these facilities has led to increased demand and wages for skilled labor, higher prices for materials like concrete and steel, and intense competition for land, especially in areas with robust utility infrastructure. For example, wholesale electricity prices near data center clusters have more than doubled since 2020, and residential electricity prices in the U.S. saw an approximate 5% annual increase in 2025, nearly double the general inflation rate.
The investment surge by major hyperscalers—Alphabet, Amazon, Meta, and Microsoft—with cumulative AI-related capital expenditure projected at $650 billion for this year, represents a 67% increase from 2025. This massive spending is creating supply-chain strains and higher input costs for components like memory chips, which saw prices jump 246% from early 2025 to December. The demand for these resources is so high that some states, including New York, Maine, Oklahoma, and Georgia, have considered or implemented restrictions on large-scale data center development. Federal Reserve Chair Jerome Powell has acknowledged that this AI-fueled data center boom is "probably pushing inflation up," challenging the idea that AI's productivity gains would immediately lower prices.
Policymakers are grappling with the fact that these inflationary pressures from AI infrastructure development are emerging faster than the anticipated productivity gains from the technology itself. This creates a challenging scenario where the economy could appear to be cooling in some consumer sectors, but remain hot in key industrial areas due to investment-led inflation. The impact spans various sectors, affecting not only technology but also utilities, power developers, and raw material suppliers like copper miners, highlighting a "K-shaped" inflation where some areas experience significant price increases while others may not.
