People naturally imagine AI as a tool that:
Helps companies get work done faster.
The same number of people can produce more.
Business costs fall.
Eventually, goods and services become cheaper.
Following this logic,
AI should be good news in the fight against inflation.
But recently, Swiss National Bank board member Petra Tschudin reminded us:
In the short term, AI could actually increase inflationary pressures.
This may sound contradictory.
The key lies in a timing gap:
People are already spending money on AI, but the promised productivity gains haven’t fully materialized yet.
Why Doesn’t AI Lower Prices Immediately Even Though It Boosts Productivity?
Let's explain "productivity" in the simplest terms.
Suppose a factory originally has:
10 people producing 100 items per day.
After implementing AI and automation,
the same 10 people can produce 120 items.
That's a productivity increase.
If all else remains equal,
the company can produce more with similar costs,
which theoretically should lower prices.
The problem is:
The real world doesn’t just wait around for AI to slowly improve productivity.
When companies see AI might change the future,
they usually start investing first.
Buying chips.
Renting computing power.
Building data centers.
Hiring engineers.
Upgrading power grids.
Buying servers.
Increasing software expenses.
Thus,
future supply hasn’t increased yet, but demand has already surged today.
Prices might rise initially.
The Swiss National Bank’s Primary Concern: Everyone Competing for the Same Limited Resources
Reuters reported on August 21 that Petra Tschudin said AI could raise inflation in the short term.
One very practical example she gave is:
chip shortages.
If many companies worldwide simultaneously ramp up AI investments,
they will compete for:
high-end chips,
memory,
servers,
data center capacity,
and other limited technological resources.
Demand spikes rapidly,
but supply can't double overnight.
Prices naturally may go up.
This is a timing issue separate from whether AI eventually improves efficiency.
Like Building a Subway: Before It Opens, Demand for Cement and Workers Drives Up Costs
The simplest analogy is subway construction.
A city government announces:
A new subway system will make transportation more efficient in five years.
In the long term,
this may be true.
Commute times shorten.
Stores get better logistics.
Companies find it easier to hire workers.
The whole city becomes more efficient.
But before the subway opens,
the first thing that happens is:
competition for workers.
competition for cement.
competition for steel.
competition for construction equipment.
competition for land.
So,
efficiency hasn’t improved yet, but construction costs rise first.
AI might be in a similar phase now.
We are building a huge "AI subway."
Electricity Could Be the Next Major Bottleneck
AI isn’t just a chat window.
Behind it lies:
data centers,
GPUs,
cooling systems,
high-speed networks,
and significant electricity consumption.
Philip Lane, ECB executive board member, pointed out in discussions about AI and monetary policy that:
Rapid growth in AI computing power increases energy demand,
and before the energy supply catches up,
this can put upward pressure on energy prices.
This logic is crucial.
If AI helps a company reduce labor costs by 10%,
but at the same time, the economy faces higher costs for electricity,
servers,
chips, and infrastructure from AI data centers,
then the overall impact on prices may not be just:
"AI saves money for everyone."
Instead, two forces pull simultaneously.
One lowers costs,
the other raises new costs.
Another Odd Effect: Belief in a Wealthier AI Future Might Increase Inflation Now
There’s another less intuitive reason.
Suppose everyone truly believes:
AI will dramatically boost economic productivity in the future.
Companies feel:
"We’ll make more profit in the future."
So, they invest now.
Households feel:
"Our incomes might rise."
So, they spend more today.
What happens then?
Today’s demand increases first.
But how much AI actually increases production
might take years to show fully.
The Bank of England’s latest research highlights:
Productivity gains don’t automatically mean inflation falls.
If expectations of productivity improvements stimulate current investment and consumption,
but actual supply hasn’t caught up,
inflationary pressures can rise instead.
In other words:
Believing we’ll be richer tomorrow can make today more expensive.
Price Drops and Inflation Declines Are Not the Same Thing
Here’s a concept that’s easy to confuse.
Assume AI reduces the price of a service from:
100 units
to 90 units.
This is a price reduction.
But inflation measures:
how fast overall prices rise.
If AI causes a one-time price drop,
then prices resume normal growth rates,
AI doesn’t necessarily keep inflation low forever.
The Bank of England research explicitly distinguishes:
one-time productivity gains
from sustained long-term productivity growth.
A one-time efficiency boost can lower the price level once,
but sustained inflation changes require longer-term, ongoing economic shifts.
So, when you hear:
"AI lowers costs,"
don't immediately conclude:
"Central banks won’t worry about inflation anymore."
There is still a long journey ahead.
This Explains Why Central Banks Are Taking AI Seriously
Previously, when central banks talked about AI,
people might have thought it was about:
whether banks should use ChatGPT.
Now, it’s much more complex.
AI is already impacting what central banks care about:
productivity,
corporate investment,
consumption,
wages,
energy demand,
chip prices,
financial markets,
and most importantly:
inflation.
What’s tricky is AI affects both supply and demand.
If AI helps companies produce more,
that lowers inflationary pressure.
But if everyone invests heavily in AI simultaneously,
demand rises.
The direction inflation eventually takes depends on:
which side moves faster.
AI Could Make Interest Rate Decisions Harder
This leads to a challenge familiar to everyone:
interest rates.
Suppose the economy suddenly overheats.
Businesses invest heavily.
Households spend a lot.
Prices go up.
In this usual case,
central banks might raise interest rates
to cool demand.
But what if this boom is driven by AI raising future productivity?
Central banks then face a tough judgment:
Is this:
real sustainable economic growth?
Or:
just spending driven by hopes for AI’s future?
Getting it wrong could:
stifle genuine productivity growth,
or let inflation rise again.
The real challenge for central banks is not:
"Will AI raise or lower inflation?"
But:
AI might do both—just at different times.
Short-Term Inflation Rises and Long-Term Declines Are Not Contradictory
The Swiss National Bank is not saying:
"AI will cause permanent inflation."
On the contrary,
Petra Tschudin also notes,
if AI sustainably boosts productivity,
it could reduce price pressures in the long run.
The true uncertainty is:
how large is this effect?
when will it appear?
how quickly does supply grow?
how much will investment-driven demand increase?
This is what central banks are watching closely now.
So the fuller AI economic story is not:
AI → higher productivity → cheaper goods.
It might be:
heavy AI investment first
↓
higher demand for chips, power, data centers, and talent
↓
some costs rise initially
↓
gradual emergence of real AI productivity gains
↓
companies produce more with the same resources
↓
long-term potential for strong cost reductions
This transition period
could be quite long.
Next Time You Hear "AI Will Fight Inflation," Ask This Question First
In the future, we’ll see many predictions:
AI lowers company costs.
AI raises GDP.
AI boosts worker productivity.
These may all be true.
But we should also ask:
How much are we spending today to gain these efficiencies?
If the whole world is simultaneously:
building data centers,
competing for electricity,
buying GPUs,
expanding power grids,
and hiring AI talent,
then we might be experiencing a fascinating phase:
AI promises a cheaper economy long-term, but building AI is expensive in the short term.
This is not AI failing.
It’s nearly every major technological revolution:
transitioning from invention
to truly raising societal productivity
requires paying a huge upfront infrastructure cost.
So the real question about AI and prices isn’t whether they go up or down today,
but:
when will new supply catch up with the new demand driven by AI investments?
Today, improve yourself alongside AI.
Learn one AI skill daily.
Save a little time each day.
Boost your abilities step by step.
SasaDaily, growing with you.
Recommended Reading
Is AI Actually Creating Value? Why UK GDP Is Starting to Show Answers
Tech Companies Promise to Cut Emissions, But 60 AI Data Centers May Emit As Much As 27 Coal Plants
If AI Saves You Time at Work, Why Do Bosses Expect You to Do More?
