Macro Markets and Machines_The Economic and Market Transformation Driven by AI_GWM report - Flipbook - Page 36
Macro, Markets, and Machines
November 2025
Section 4.7: Inflationary Impact and Policy Response Is Uncertain
The inflationary impact of AI remains uncertain, but estimates are biased higher across the existing research.
This is despite positive productivity shocks generally thought of as disinflationary, as greater per worker output
and, in turn, overall output, all else equal, results in a higher supply of goods and services in the economy,
prompting lower prices or fewer price increases. With AI shocks characterized by uncertain degrees of worker
complementarities against worker displacements, it is difficult to make assumptions on how AI may affect the
labour force and, in turn, household incomes that are a key determinant of aggregate demand and therefore
consumer prices.
The IMF’s model incorporates a forward-looking component that generates only a modest inflationary
response above its baseline projections. Expectations of higher productivity and thus higher future incomes,
combined with present-day income gains – concurrent with growing AI benefits or greater investment and hiring
to capitalize on the technology – cause an increase in aggregate demand that triggers higher inflation in the near
to medium term – though broad-based labour dislocation and wage suppression could work in the opposite
direction. The rapid infrastructure build-out and demands of increased AI adoption in the workplace may
already be evident in economic data.
Along these lines, the International Energy Agency (IEA) projects in its Energy and AI report that electricity
demand from global data centres will more than double over the next five years, with AI-optimized data
centre demand expected to quadruple over the period (Exhibit 29). The IEA’s report stipulates that datacentre
power consumption will represent half of the domestic growth in energy demand projected by 2030, with the
sector’s energy usage overtaking the combined demands of all energy-intensive-goods sectors, including
aluminum, steel, cement, and chemicals manufacturing.
All else equal, this would of course pressure global energy prices higher, although greater investments in
energy production (renewable or not) or supply policies (e.g., OPEC+) can materially influence the medium-term
direction of energy prices. In parallel to higher power demand, the AI rush could also affect the prices of
electronic devices such as personal computers and cell phones given the emergence of higher business demand
for advanced chips. The concentration of energy and critical mineral resources adds geopolitical risk that could
compound cost pressures (with implications for the AI access shortcomings described in Section 4.4).
Exhibit 29 – Data Centre Power Demand Is Expected to Surge
1000
900
800
700
TWh
600
500
400
300
200
100
0
2020
Other infrastructure
2022
Cooling
2024
Other IT equipment
2026
Conventional servers
2028
2030
Accelerated servers
Sources: International Energy Agency; Scotia Wealth Management.
Scotia Wealth Management
35
