IMF Economist Warns AI Infrastructure Boom Could Fuel Inflation

The long-held hope that artificial intelligence will act as a powerful disinflationary force is facing a critical challenge. Silvana Tenreyro, an economist at the IMF, warns that AI-driven productivity gains may not necessarily lower prices and could, in some scenarios, actually drive inflation higher.

In research published on the Bank of England's 'Bank Underground' blog, Tenreyro and co-authors Jenny Chan and Ludovica Ambrosino argue that the inflationary impact of AI is ambiguous. The authors highlight a specific risk: the gap between investment and realization. They note that business investment and household spending can move ahead of realized productivity gains. This phenomenon is currently occurring with the massive build-out of AI infrastructure, such as data centers. When investment demand surges before the economy sees actual efficiency gains, it can create supply crunches that push prices upward.

The Sector Divide

The research explains that the effect on inflation depends heavily on where productivity gains occur. According to the authors, productivity gains within the services sector typically have a disinflationary effect, lowering domestic prices. However, the opposite can happen when gains are concentrated in exports. In those cases, increased productivity in export sectors can drive up domestic wages and heighten demand for services that are already supply-constrained, ultimately raising overall inflation.

Challenging the Optimists

These findings challenge the optimistic outlook held by some policymakers. Federal Reserve Chair Kevin Warsh has argued that AI will be a significant disinflationary force, potentially allowing for faster economic growth without the traditional risk of overheating the economy. Tenreyro’s analysis suggests this may be an oversimplification. If the transition to an AI-driven economy is characterized by front-loaded infrastructure costs and supply bottlenecks, the path to higher productivity may be volatile.

Market Implications

If the analysis by Tenreyro and her colleagues is correct, the transition to an AI-integrated economy could have significant implications for monetary policy. Rather than providing a cushion that allows for lower interest rates, the temporary inflation spikes caused by infrastructure investment might necessitate higher interest rates to keep prices stable. This creates a paradox where the very technology intended to boost long-term efficiency could lead to short-term financial tightening, complicating the mandate of central banks as they navigate the AI transition.