Guggenheim's Schwartz: Power Crunch Could Cost U.S. AI Leadership

Alan Schwartz, Executive Chair of Guggenheim Partners, told attendees at the Milken Institute Global Conference that growing electricity demand from data centers and artificial intelligence (AI) workloads is exposing critical gaps in the U.S. power system. He warned that if transmission, permitting and local capacity shortfalls are not addressed, the United States risks losing momentum in AI development to better-prepared competitors.

Schwartz outlined how the issue unfolds in practice: hyperscale compute clusters require rapid access to large, reliable blocks of power and long interconnection lead times and strained local distribution networks can prevent completed facilities from becoming fully operational. Panel discussion with energy and data center leaders emphasized near-term mitigation options — such as on-site generation and accelerated grid upgrades — while noting potential trade-offs with decarbonization goals.

Market implications are already visible according to several reports: major banks and industry studies project substantial increases in data-center power demand over the coming years, which could tighten reserve margins and influence utility and infrastructure equities. Investors should monitor regional capacity forecasts and utility rate cases as potential drivers of sector performance.

In a broader context, policymakers and technology leaders have previously warned lawmakers that aging infrastructure and slow permitting risk constraining U.S. AI scale-up, a concern echoed during recent congressional hearings on competitiveness. The panel underscored calls for streamlined approvals, targeted transmission investments and public-private collaboration to keep the domestic AI ecosystem competitive.

Looking ahead, analysts expect short-term operational constraints in high-demand regions and an investment cycle in grid reinforcement and long-haul transmission. For market participants, near-term catalysts will include government policy moves on permitting, large cloud providers’ energy strategies and quarterly updates from utilities and major data-center operators. Absent coordinated action, the power bottleneck could become a structural headwind for U.S. AI ambitions.