Help World Market Atlas grow
Ordering a company valuation also supports the project’s development.
Support the project
Help develop World Market Atlas — an independent project for individual investors. Your support helps us produce analyses and improve our tools.
Market commentary
My perspective on markets, the economy and events that matter to investors.
Inflation Will Stay Until the AI Bubble Bursts
A shortened analysis drawing on academic literature and institutional sources. Sources reviewed: 12 October 2026. The title presents a hypothesis, not a certain forecast.
Artificial intelligence may raise costs before it helps reduce them. Building data centres increases today's demand for electricity, equipment and workers. Efficiency gains take time. That supports concern about slower disinflation, but does not establish that only a bursting bubble can end it.
I treat the title as a hypothesis. High valuations and rapidly rising investment do not establish that AI is a bubble. Here, persistent inflation means price growth above the central bank's target, rather than simply positive inflation.
1. Investment now, savings later
In the short run, firms buy servers, construct facilities and hire workers. Prices can rise if resource supply cannot keep up. On 28 September 2026, Federal Reserve Governor Lisa Cook assessed that AI investment was adding inflationary pressure and delaying inflation's return to the 2% target. This is an institutional assessment, not a measurement of AI's contribution to inflation. [1]
The productivity J-curve literature explains why benefits may arrive later: new technologies require training, organisational changes and complementary investment. Purchasing equipment alone does not guarantee an immediate increase in economy-wide efficiency. [2]
2. Electricity can become a bottleneck
The IEA estimates that data centres consumed approximately 485 TWh of electricity in 2025. Its projection reaches 950 TWh in 2030, or around 3% of global electricity demand. These figures cover all data centres, not AI alone. [3]
Pressure may be intense where grid connections and generating capacity are scarce. However, an increase in electricity demand cannot be translated into an identical increase in consumer inflation: supply availability, tariffs and cost pass-through matter.
3. Productivity is a real counterargument
Brynjolfsson, Li and Raymond studied the introduction of an AI assistant among 5172 customer-support workers. Average issues resolved per hour increased by 15%. This is evidence from one application, not proof of an equivalent productivity gain across the economy. [4]
If similar benefits spread and savings reach customers, AI could reduce price pressure. Inflation could therefore fall through greater supply and efficiency while investment continues. Adoption speed and competition will influence how much of the benefit remains in company margins.
4. What would a bursting bubble change?
Scenario: disappointing revenue lowers valuations, weakens access to finance and leads to project cancellations. Lower investment reduces demand for equipment, construction and electricity. Lost stock-market wealth may also weaken consumption. Research documents the stock-market wealth effect; applying it to a future AI crash remains an interpretation. [5]
This process could lower inflation at the cost of economic activity. A stock-market correction alone is insufficient if investment and infrastructure use continue. Moreover, reduced investment could weaken future supply, while an independent energy shock could sustain price increases.
5. Implications for investors
The most defensible proposition is that the AI boom may prolong above-target inflation while its demand impulse exceeds its supply benefits. A crash is one possible route to weakening that impulse. It is neither a necessary condition nor a guarantee that inflation returns to target.
I would monitor actual capital spending, AI revenue, electricity costs, productivity and central-bank decisions. Without an original causal study, I do not assign AI a specific number of percentage points of inflation or a date for its normalisation.
Sources
Posts reflect the author’s personal views as of publication. They are not investment recommendations.