Citadel Securities: AI Theme Faces Rising Political Pressure, Inflation Risks May Force the Fed to Hike Rates
Taylor Wilson
Citadel Securities analyst Shah warns that AI infrastructure spending, tightening energy markets, and a firming labor market are jointly pushing inflation higher — the Fed's next move is more likely a hike than a cut, and tighter financial conditions are the market's next major risk.
How much money is AI actually burning?
AI capital-expenditure commitments for 2026 alone exceed $700 billion, yet compute supply still cannot keep pace with increasingly powerful, compute-hungry models.
AI is no longer just a capex story. Anthropic and peers are posting rapid annualized revenue growth — enterprise AI adoption is converting into real economic demand.
This means → the spending is not idle. It simultaneously pulls on energy, labor, and construction — inflation pressure is arriving from multiple directions at once.
Where is the inflation pressure coming from?
Energy: the Strait of Hormuz blockade has shocked supply chains. Even after the conflict ends, firms and governments will spend more on energy security, hold larger inventories, and diversify supply — reversing pre-crisis efficiency gains and raising energy, transport, insurance, and capital-investment costs.
Labor: June nonfarm payrolls added 172,000 jobs; unemployment remains low, layoffs are historically suppressed, and tighter immigration policy is shrinking supply. In plain terms = the labor market is a taut string — once AI investment and fiscal stimulus accelerate together, wage growth could overshoot the Fed's tolerance, keeping services inflation elevated.
Data-center bottlenecks: power supply, grid access, transformers, construction labor, cooling equipment, specialist engineering capacity — AI expansion has hit not one wall but a row of walls. Demand exceeding the industry's ability to scale is pushing costs up across the entire data-center value chain and spilling into broader construction, industrial, and energy markets.
Could the Fed really hike?
The Dallas Fed's high-frequency GDP tracker puts U.S. growth at an annualized 3.23% — resilience that continues to beat expectations.
Fed Governor Waller has shifted from supporting cuts to advocating the abandonment of an easing bias; Dallas Fed President Logan has openly discussed the possibility of further hikes.
This means → the internal winds at the Fed have already turned. Shah's bottom line: robust AI-driven investment, tighter energy markets, an improving labor market, and infrastructure bottlenecks all point to upside risk for nominal growth and inflation. "The Fed's next move is most likely a hike — and perhaps soon."
Is AI a bubble? What does Shah say?
Shah characterizes the current AI market as "narrative frenzy," not "bubble" — the core driver behind surging capex revisions is that compute supply remains insufficient.
He flags that markets focus too heavily on what compute demand means for semis and infrastructure spending, and too little on what it means for adoption economics — the question of whether enterprises actually earn a return on AI.
In plain terms = more powerful models create more value but also consume more compute. Whether AI can sustain itself comes down to one question: can productivity gains outrun the cost of producing those gains?
What does this mean for markets?
Political scrutiny of AI is intensifying, adding extra uncertainty to the process of validating AI's commercial returns.
Shah's core warning: tighter financial conditions are the market's next major risk.
This reflects a deeper contradiction — the more successful AI is at pulling real-economy demand, the greater the inflation pressure, the stronger the Fed's case for tightening, and the more that tightening undermines the valuation foundation of AI investment itself.
Content is for reference only, not financial advice.