The Fed Confronts the Price of Intelligence

“Capital is being repriced before the productivity dividend has been fully realised.”

THE BIG TAKE — The most consequential new signal is that the Federal Reserve is beginning to confront the inflationary consequences of AI capital formation alongside an emerging divergence in employment income and financing conditions.

In an October 30 speech delivered after the previous monetary-policy assessment, St. Louis Federal Reserve President Alberto Musalem presented an unusually explicit diagnosis.

AI-related capital expenditure continues to support strong American economic growth. Yet financing costs are rising for some AI infrastructure borrowers, specialised construction wages are accelerating, and broader real household incomes remain under pressure.

Musalem concluded that additional monetary tightening will be necessary soon. His remarks reinforce the hawkish message delivered by Dallas Fed President Lorie Logan on the same day, although neither statement constitutes a new FOMC decision.

Musalem identifies three developments that deserve particular attention.

First, data-centre construction is supporting double-digit growth in nonresidential fixed investment.

Second, real wages are rising rapidly in specialised occupations, including electricians and skilled construction workers serving the data-centre industry.

Third, credit spreads have widened, particularly among higher-risk companies financing AI infrastructure.

Meanwhile, nominal wage growth across the broader economy has failed to keep pace with inflation.

The result is an emerging divergence between the sectors building the AI economy and households that have not yet captured its productivity benefits. This is an important dimension of the Great Split.

The investment cycle is generating concentrated demand for capital, infrastructure and specialised labour before its benefits become broadly distributed.

Musalem outlined two alternative scenarios over an 18-month horizon.

His more likely scenario involves continued demand strength and persistent supply constraints, requiring additional increases in interest rates to return inflation toward 2%.

His alternative scenario involves stronger productivity growth, falling energy costs or weaker investment demand, potentially allowing inflation to moderate without additional tightening.

He explicitly includes a substantial slowdown in AI and data-centre spending among the developments that could reduce inflationary pressure.

The distinction is critical. Productivity-driven disinflation would be economically constructive. Disinflation caused by collapsing investment expectations would not necessarily be.

Musalem considers the latter combination of disinflationary forces less likely than continued inflation persistence.

A particularly revealing finding concerns financing conditions.

Musalem observes that AI infrastructure borrowers are already experiencing wider credit spreads. This introduces a potentially self-reinforcing financial mechanism.

HIGHER POLICY RATES → HIGHER FINANCING COSTS → GREATER SELECTIVITY IN AI LENDING → SLOWER CAPITAL FORMATION → DELAYED PRODUCTIVITY RETURNS.

At the same time, disappointing investment returns could produce a broader revaluation of technology assets.

The distinction between these two risks matters. A gradual repricing of AI credit could improve capital discipline. A disorderly withdrawal of financing could damage productive investment and transmit instability into other financial markets.

For European investors, this makes the composition of AI financing increasingly important: corporate cash-flow generation, debt maturities, credit quality and dependence on refinancing deserve greater attention than headline capital-expenditure announcements alone.

The United States is confronting the inflationary consequences of extraordinary technological investment. Europe faces an additional challenge: financing the diffusion of AI throughout its productive economy.

If American investment demand continues supporting higher global real interest rates, European companies could face more expensive capital without experiencing comparable domestic productivity growth. That would reinforce the Great Split.

The ECB must therefore distinguish the imported financing consequences of the American AI boom from Europe's own underlying productive investment requirements.

The appropriate response is not preferential monetary treatment for AI projects. It is a financial architecture capable of allocating capital efficiently across infrastructure, industrial adoption, organisational innovation and workforce development.

Five Takeaways For Investors And EU Policymakers

the actionable read

• The Federal Reserve's hawkish position is broadening. Musalem's intervention reinforces Logan's October 30 call for additional tightening, although the FOMC has made no further decision.
• The labour-market divergence deserves closer attention. Rising specialist construction wages alongside weak broader real wage growth illustrate how investment concentration can distribute the benefits and costs of technological change unevenly.
• AI credit spreads are becoming a leading indicator. Higher borrowing costs may expose the difference between financially sustainable infrastructure projects and investments dependent on optimistic future valuations.
• Industrial-AI deflation remains conditional. Lower unit production costs must become visible across the productive economy before central banks can confidently incorporate a significant AI dividend into their inflation outlook.
• Europe's strategic challenge is capital productivity. EU policymakers should prioritise investment diffusion, financing of intangible assets and competitive energy infrastructure while strengthening supervision of concentrated AI-related credit exposures.

The latest Federal Reserve intervention strengthens an important conclusion. The AI Supercycle is creating immediate demand for capital, electricity, construction and specialised labour. Its potential productivity benefits remain uneven and uncertain.

Financial markets are simultaneously becoming more selective about financing the infrastructure required to deliver those benefits.

For investors, the next market rotation will be determined by the relationship between realised productivity, sustainable cash flows and the rising cost of capital—not simply by the scale of AI investment.

For Europe, the fundamental danger is that international financing costs rise faster than domestic productive capacity.

The Great Split is becoming a financial reality: capital is being repriced before the productivity dividend has been fully realised.