Central banks are racing to understand how AI is reshaping inflation — and they do not all agree on the answer.
For most of the past three years, central banks have treated artificial intelligence the way they treat climate change or demographics: a long-horizon force worth monitoring, but not yet an instrument of monetary policy.
ADVERTISEMENT
ADVERTISEMENT
That distinction has now collapsed.
Policymakers have begun to frame AI as a structural shift on the scale of electrification or the internet, one that will reshape inflation, interest rates and the very tools central banks use to set them.
The disagreement is no longer about whether AI matters.
It is about timing, transmission and direction: how quickly its effects will materialise, whether prices rise or fall first and how central banks should respond to a force that may be inflationary in the short term but disinflationary over time.
What the ECB and Bundesbank are already doing with AI
The European Central Bank has moved fastest in turning AI from theory into practice.
In a blog post published on 21 April 2026, four ECB economists — Óscar Arce, Karin Klieber, Michele Lenza and Joan Paredes — disclosed that since the end of 2022 a machine learning model has been part of the analytical toolkit used to prepare monetary policy decisions for the Governing Council.
The model draws on roughly 60 indicators capturing inflation expectations, cost pressures, real economic activity and financial conditions, and is updated several times each quarter.
The results have already been tested in real time.
In the second and fourth quarters of 2025, the model flagged upside risks to core inflation that later materialized, with final readings coming about 20 basis points above official Eurosystem projections.
"Artificial intelligence (AI) can help track inflation risks in real time," the authors wrote.
The Bundesbank is moving along a similar path.
At a joint Bundesbank–SUERF conference in Frankfurt on 9 December 2025, Bundesbank President Joachim Nagel affirmed that the German central bank is already using a wide range of AI applications to improve analyses and support work processes.
These include text-based intelligent assistants, AI-driven document analysis and a model called MILA that evaluates communications from euro area central banks.
"Technology should ultimately serve people. And the same holds for us as central banks: we use AI to fulfil our mandate as well as possible," Nagel said.
What the Fed is saying about AI: From curiosity to core debates
At the Federal Reserve, the shift has been less operational but more conceptual and increasingly urgent.
Officials have moved from acknowledging AI to debating how it reshapes the core trade-offs of monetary policy.
Last year, Federal Reserve Governor Christopher Waller argued that AI is being adopted faster than personal computers, the internet or smartphones, and that the productivity question now sits at the centre of the monetary policy debate.
"A crucial question is whether AI will contribute to a resurgence in productivity growth. Any sustained productivity growth above 2 percent will tend to support rising real incomes and living standards without inflation pressure. As a monetary policymaker, I'm hoping that AI delivers," Waller said.
Speaking at the Euro20+ event hosted by Nagel in November 2025, Federal Reserve Vice Chair Philip Jefferson highlighted AI's double-edged effect on inflation.
On one side, the technology could lower production costs through productivity gains. On the other, it could push input prices higher.
"AI could put upward pressure on certain price categories as many firms push to scale up the technology. AI technology also requires data centers, which compete with other production processes for land, energy, and other inputs. So, I think that AI's effect on inflation is not solely downward pressure," Jefferson said.
The most politically charged voice in the debate is Kevin Warsh, nominated by Donald Trump to chair the Fed when Jerome Powell’s term expires in May.
Warsh has called the AI boom the most productivity-enhancing wave of his lifetime and compared the moment to the late 1990s, when Alan Greenspan kept policy looser than mechanical rules suggested and was rewarded with rising productivity and stable prices.
But his confirmation hearing testimony last week showed a more cautious side.
He described AI as a force approaching "escape velocity" and warned policymakers cannot yet rely on those productivity gains.
“AI is becoming so consequential that it is approaching something like escape velocity,” Warsh said, warning that the Federal Reserve may need to rethink its models.
While he acknowledged that the current wave of innovation could, over time, ease price pressures and simplify the inflation fight, he cautioned that policymakers still lack clarity on how those gains will feed through to employment — the other half of the Fed’s mandate.
Wall Street is split between disinflation bulls and capex hawks
While central banks are still debating how to interpret AI, Wall Street has begun trading on it. The largest asset managers and investment-bank economists are pricing the technology into their forecasts for inflation, growth and bond yields — and the Street has split into two opposing trades.
The disinflation bulls treat AI as a positive supply shock: lower prices, lower rates, higher risk assets.
The capex hawks see a near-term inflation problem: a record investment cycle that pushes up electricity prices, drains the savings-investment balance and lifts long-end yields before any productivity gains arrive.
The most aggressive disinflation case has come from the asset management industry. Mike Hunstad, head of Northern Trust's $1.4 trillion asset management division, told the Financial Times in April 2026 that AI could prove to be one of the biggest positive supply shocks in modern economic history.
According to Hunstad, if AI delivers a sustained productivity uplift, it will do the disinflationary work that years of restrictive policy could not finish.
"It's almost like AI is your monetary policy, and it's going to be more effective than anything the Fed or really any central bank around the world can do," Hunstad said.
The capex hawks see the opposite: an investment cycle large enough to lift electricity prices and yields before any productivity gains arrive.
Oxford Economics' Ben May and Daniel Harenberg argued in February that pre-emptively cutting rates on the assumption AI will be disinflationary would be a mistake — AI is currently boosting inflation through electricity prices, data centre investment and wealth effects from rising stock prices.
"AI's impact on inflation will rest on the extent to which the boost to the supply side of the economy is offset by any associated rise in aggregate demand," Oxford Economics stated in its note.
Goldman Sachs reached a similar conclusion. Economists Manuel Abecasis and Hongcen Wei, focused on what may be the most underappreciated transmission channel from AI to inflation: electricity prices.
Power inflation in the US ran at 6.9% year-on-year through December 2025, well above headline PCE inflation of 2.9%. Goldman expects consumer electricity inflation to remain near 6% in both 2026 and 2027 before decelerating to about 3.5% in 2028.
"We expect data centers to boost electricity demand significantly, accounting for about 40% of total power demand growth over the next five years," Abecasis said.
Goldman estimates higher electricity costs will add 0.2 percentage points to headline inflation in 2026 and 0.15pp in 2027, with the medical, transportation and food services sectors absorbing the bulk of the indirect pass-through into core prices.
The question is timing, not destination
There is now broad agreement on one point: AI is large enough to force central banks to rethink how the economy works.
What remains unresolved is sequencing. If productivity gains arrive first, central banks may gain room to cut rates without reigniting inflation.
If the investment boom hits first — through energy prices, capital demand and asset valuations — policymakers who ease too early may be forced into a reversal.
That marks a sharp break from just a few years ago, when artificial intelligence barely appeared in central bank speeches.
