Castle Securities warns: AI may initially drive up inflation before bringing productivity gains.
```
The impact of AI on the macroeconomy may not necessarily follow the familiar "productivity boost—inflation decline" path anticipated by the market. Citadel Securities points out that before efficiency dividends are realized, the AI transformation may first raise costs through investments, training, and organizational restructuring, making inflation, growth, and interest rate outlooks more complex.
In the short term, the easing of energy shocks has become the main driver of macro markets. Citadel Securities notes that, influenced by progress in US-Iran negotiations, crude oil futures have dropped about 12% recently. The firm estimates that if the Strait of Hormuz is fully reopened by the end of July, US 10-year yields could fall by about 12.5 basis points, the S&P 500 could rise by 1.75%, and the broad dollar might fall by 0.5%.
The more critical medium-term issue is that policy makers are simultaneously facing generational transformations driven by AI and significant supply-side shocks. Citadel Securities warns that, in a context where tariffs, geopolitical conflicts, and limited labor supply have already raised the starting point for inflation, if AI capital expenditure and economic growth continue to strengthen, inflation could once again become the central risk most needing market attention.
Cyclical vs Structural: How AI Disrupts Macro Indicators
Citadel Securities notes that the current core difficulty for policymakers lies in distinguishing between cyclical and structural changes. Traditional narratives believe that AI will enhance productivity, lower costs, ease inflation pressures, and support stronger growth.
But Nohshad Shah argues that AI is more likely to act as an inflationary force before it becomes a deflationary one—and this process may occur before measurable gains in productivity appear. Reasons include initial costs for AI implementation, employee retraining expenses, and disruptions from organizational adjustments.
He cites a recent paper by Simone Lenzu of the New York Fed which offers a similar viewpoint: AI blurs the traditional boundaries between cyclical fluctuations and structural changes, making standard macro indicators harder to interpret. The AI transformation may change labor demand, production processes, and pricing behavior, weakening the link between economic slack and inflation.
This means that traditional indicators like the unemployment rate and output gap may no longer be as reliable as before. For central banks, judging whether the economy is overheating and whether policies are sufficiently restrictive will become more complicated.
Disconnect Between Expectations and Reality: Markets Race Ahead of Productivity Dividends
Citadel Securities points out that another issue brought by AI is that the market may realize productivity expectations ahead of the real economy. Expectations of future efficiency improvements can boost investment, consumption, and asset prices prematurely, while actual productivity improvements are yet to manifest, widening the gap between market expectations and economic reality. Nohshad Shah says this phenomenon may already be visible in the current stock market.
Even if AI ultimately delivers significant productivity gains, it may raise the economy’s potential growth rate and equilibrium real interest rate, forcing central banks to reassess the neutral rate. In other words, AI is not only a growth variable—it will also impact inflation dynamics, interest rate paths, and financial stability.
Citadel Securities concludes that the current stage of AI transformation may form a tricky combination: high inflation, weak measurable productivity, and elevated asset prices. This combination is challenging for both policy-making and market pricing.
Risk Warning and DisclaimerMarkets are risky; investment requires caution. This article does not constitute personal investment advice, nor does it take into account the specific investment goals, financial situation, or needs of any individual user. Users should consider whether any opinions, views, or conclusions in this article are suitable for their particular circumstances. Investing based on this article is at your own risk. ```