NextFin News - Federal Reserve Chair Jerome Powell has upended the prevailing economic narrative that artificial intelligence would serve as an immediate deflationary force, warning instead that the massive physical buildout of AI data centers is actively pushing inflation higher. Speaking at a press conference following the Federal Open Market Committee’s decision to hold interest rates steady on March 18, 2026, Powell identified the "insatiable" demand for data center infrastructure as a primary driver of price pressure across the American economy. The admission marks a significant pivot for the central bank, which had previously focused on the long-term productivity gains of AI rather than the immediate inflationary costs of its construction.
The scale of the infrastructure boom is staggering. U.S. President Trump’s administration has prioritized American technological dominance, but the resulting race to build "data centers everywhere," as Powell described it, has created severe bottlenecks in the supply of copper, specialized semiconductors, and electrical power. This surge in demand is not just a tech-sector phenomenon; it is bleeding into the broader economy by driving up electricity rates for households and competing for the same raw materials used in traditional manufacturing and housing. Powell noted that in the short term, the demand side of the AI revolution is running far ahead of any productivity payoff, effectively raising the "neutral" interest rate and complicating the Fed’s path toward further easing.
For investors, this shift necessitates a departure from the "growth at any cost" AI playbook that dominated 2024 and 2025. The new environment favors companies with genuine pricing power—those capable of passing on rising input costs without sacrificing volume. Freeport-McMoRan has emerged as a central figure in this landscape, as its copper production becomes the literal nervous system of the AI buildout. With the company estimating it could increase output by 60% by 2030, it stands as a primary beneficiary of the very inflationary pressures Powell is tracking. The scarcity of copper, essential for the dense wiring required in high-density server racks, ensures that miners hold the upper hand in price negotiations.
The bottleneck strategy also extends to the energy sector, where the "nuclear renaissance" has moved from theory to a critical market driver. Constellation Energy, now the world’s largest private-sector power producer following its merger with Calpine earlier this year, is uniquely positioned to capitalize on the tech giants' desperation for carbon-free, 24/7 electricity. As AI models grow more power-hungry, the ability to provide reliable baseload power has become a more valuable commodity than the software itself. This represents a fundamental re-rating of the utility sector, transforming once-stodgy dividend plays into high-growth infrastructure assets essential to the national interest.
In the semiconductor space, the focus is shifting from the logic chips designed by Nvidia to the memory bottlenecks that currently constrain AI performance. Micron Technology, the only major U.S.-based supplier of high-bandwidth memory (HBM), has already sold out its entire 2026 supply. Management’s recent admission that they can only meet roughly half of the demand from key customers underscores a supply-demand imbalance that is inherently inflationary. While the broader market remains fixated on software applications, the real "inflationary AI trade" is found in the physical constraints of the physical world—the chips, the cables, and the kilovolt-amps that make the digital revolution possible.
Powell’s commentary suggests that the Federal Reserve is prepared to tolerate higher-for-longer interest rates if the AI buildout continues to overheat the industrial economy. While some economists, including U.S. President Trump’s nominee for the next Fed chair, Kevin Warsh, have argued that AI will eventually be a disinflationary force, Powell’s focus remains on the "here and now" of construction and consumption. The central bank’s cautious stance implies that the productivity miracle promised by AI may be delayed by the very costs required to build it, leaving investors to navigate a landscape where the most profitable AI companies are those that have never written a single line of code.
Explore more exclusive insights at nextfin.ai.
