States impose new barriers to speculative US data centers

Solmar Insights

Efforts to tighten data center development practices in the US are accelerating as utilities and state policymakers respond to growing public opposition and concerns over speculative project requests. Since January 2024, an estimated $170 billion of AI-focused data center capacity has been blocked, withdrawn, or stalled by community resistance and stricter regulatory scrutiny, complicating investment decisions for institutional buyers and developers.

Public resistance drives policy shifts

Community opposition has emerged as a powerful obstacle to new data center builds, particularly those focused on artificial intelligence. Citizens in regions such as Mason, Michigan and Abilene, Texas have organized high-profile campaigns at local government meetings, protesting the large amounts of energy and land resources these projects require.

The backlash is increasingly translating into regulatory responses. States are considering or enacting stricter development standards, and in some cases, moratoriums on new data center facilities. These interventions reflect a broader reassessment of how much new power-intensive digital infrastructure local grids and communities can absorb.

According to advisory firm Relae, community resistance alone has either stalled or entirely prevented projects representing over $170 billion in total planned capacity since the start of 2024. This intervention represents a sizable portion of the $581 billion in US AI infrastructure investment Goldman Sachs forecasted for this year.

Utility responses to speculative requests

Utilities, facing interconnection queues crowded with data center proposals, are rethinking their approach to new large-load requests. Historically, many requests have never materialized into actual builds, consuming significant planning resources and contributing to grid uncertainty.

To clamp down on speculation, utilities are introducing new tariffs for large loads and applying more rigorous vetting processes. These mechanisms aim to weed out noncommittal developers and projects without clear end-use cases or power procurement strategies.

The utility response is complicated, however, by the profit model for regulated utilities, which historically incentivizes the addition of load. Despite this, utility planners are increasingly factoring in risks such as strained grid capacity, investment lead times for critical equipment, and the need for community buy-in before approving new interconnection requests.

AI boom vs. project delivery bottlenecks

The AI-driven data center construction wave shows no sign of abating, with analysts predicting continued growth in project proposals and infrastructure investment. Goldman Sachs projects that US data center power demand will more than double from 2025 to reach 66 GW by 2027. Despite fast-rising demand, only about half of the capacity currently scheduled for completion in the next one to two years is expected to be delivered on time, due to ongoing challenges.

The most significant hurdles include land use conflicts, limited power availability in primary markets, construction labor shortages, and long lead times for specialized electrical equipment. Delays in project approvals and interconnection approvals compound these challenges, emphasizing the tension between market demand from hyperscalers and the ability of local stakeholders and the grid to keep pace.

By 2030, according to the Electric Power Research Institute, data centers could account for 9 to 17 percent of total US electricity demand, potentially rising to 20 percent beyond that timeframe. These figures frame the urgency with which utilities and regulators are seeking to control speculative growth.

Investor and developer implications

For institutional buyers and developers, stricter vetting and growing public scrutiny complicate project underwriting and execution. Regulatory risk surrounding power commitments, interconnection, and zoning must now be layered on top of technical, market, and financing concerns. The result is longer project timeframes and a greater requirement for pre-commitment of load and clear demonstration of community support before the groundbreaking.

With only about half of near-term capacity likely to be delivered as planned, returns on site and power acquisition may be highly variable. This environment favors well-capitalized developers who can engage early with utilities and local governments to de-risk projects, secure priority in queues, and withstand delays caused by public pushback or equipment bottlenecks.

Reporting via the original publisher