Investment can arrive before the gain
Productivity means producing more with the same inputs, which can reduce unit costs. The analysis warns that businesses and households can move first. Companies invest heavily in infrastructure and people spend in anticipation of greater future income before the economy has produced more output.
When that demand hits constrained chips, electricity, grid connections, construction capacity, transformers, and specialized labor, prices can rise. Higher interest rates may then be needed even if the long-run technology ultimately improves productive capacity.
The sector determines the inflation path
Productivity gains in domestically consumed services are more likely to reduce domestic prices directly. Gains in export sectors can instead raise wages and income, increasing demand for local services whose supply cannot expand as quickly.
Reuters notes that memory and graphics-chip prices have already risen alongside data-center demand, adding pressure to consumer electronics. That is a current supply signal, not proof of the final economy-wide inflation effect.
Do not let a forecast become a policy assumption
The research was published on the Bank Underground staff forum and does not necessarily represent an official central-bank view. It maps plausible mechanisms rather than predicting one inevitable path.
Policymakers should track realized productivity, capacity brought online, sector-specific prices, wages, demand, and the cost of capital separately. AI can be economically transformative while the transition remains inflationary, uneven, and expensive.
Go to the source
Read the evidence behind this analysis. External links open in a new tab.
Reuters — AI productivity gains may not curb inflation Bank of England — July 2026 business conditions and AI inflation pressures
