1Q26_ Quarterly Outlook Report_Final_EN - Flipbook - Page 38
T H E P LUMB LI N E | A RETU RN TO F I RS T PRI N CI PL ES
exerting downward pressure on prices. AI-driven efficiency gains have the potential to reinforce
this trend.
Labour costs have lagged corporate profits post-dotcom
2000
35000
Information Age
28000
1500
Index = 100 (1Q47)
21000
1000
14000
500
7000
0
1947
0
1958
1969
Information age
1980
1991
Unit labour costs - LHS
2002
Output - LHS
2013
2024
Corporate profits - RHS
Source: IMF, John C. McCallum (2023); U.S. Bureau of Labor Statistics (2024), Bloomberg Finance LP, Scotia Wealth Management
Technological progress also tends to compress input costs over time. An example of this can be
seen with the inflation-adjusted cost of digital storage, which has declined dramatically over the
past several decades as improvements in computing efficiency and scale effects took hold.
Similar cost curves may ultimately emerge across the AI space over time.
Technological progress causes input cost curves to collapse
$100t
$1t
US$/terabyte
$10b
$100m
$1m
$10k
$100
1959
1965
RAM
1971
1977
1983
Disk memory
1989
1995
Flash memory
2001
2007
2013
2019
Solid-state drives
Source: Our World in Data John C. McCallum (2023); U.S. Bureau of Labor Statistics (2024); OECD; Andre et al, 2025; Scotia Wealth Management.
As such, AI’s inflationary impact is best understood in two stages as short-term price pressures
driven by physical constraints are ultimately offset by longer-term disinflation once capacity
expands, diffusion accelerates, and productivity gains become widespread.
Technology is the answer to demographic challenges
By far the greatest source of anxiety surrounding artificial intelligence is the fear that it will render
human labour redundant, leading to widespread and persistent unemployment. While AI is likely
37
