ECB AI Model Hits Forecast, Bundesbank Backs AI, US Power Inflation Surges

Central banks are moving from watching AI as a distant trend to embedding it in policy analysis. The European Central Bank (ECB) has been running a machine‑learning model since late 2022 that ingests about 60 indicators—ranging from inflation expectations to financial conditions—and updates several times each quarter. In the second and fourth quarters of 2025 the model warned of upside risks to core inflation; the actual readings ended roughly 20 basis points (0.20 percentage point) above the official Eurosystem forecasts.

Bundesbank President Joachim Nagel reinforced the shift at a December 2025 conference, stating that AI tools are essential for central banks to fulfill their mandates. The German central bank already employs text‑based assistants, AI‑driven document analysis, and a model called MILA that parses communications from euro‑area banks.

Across the Atlantic, US data show a different pressure point. Power‑sector inflation rose 6.9% year‑on‑year through December 2025, far outpacing the headline personal consumption expenditures (PCE) inflation rate of 2.9%. Analysts link the surge to AI‑related demand for data‑center capacity, which competes for land, energy and other inputs, pushing input costs higher.

The convergence of these developments suggests AI is becoming a near‑term inflation driver while also promising longer‑term productivity gains. The ECB’s successful prediction demonstrates that AI can sharpen real‑time risk monitoring, a capability that may soon be standard for policy‑making bodies. Germany’s explicit endorsement signals broader institutional adoption, potentially accelerating AI‑driven efficiencies in data handling and forecasting.

In the United States, the power‑inflation spike illustrates the double‑edged nature of AI: while firms invest heavily in AI infrastructure, the resulting demand for resources can lift prices in specific sectors. Policymakers will need to balance these short‑run pressures against the anticipated productivity boost that AI could deliver over the next decade.

What to watch next: whether the Federal Reserve incorporates AI‑derived risk signals into its policy framework and how quickly AI‑related cost pressures permeate other commodity and services markets.