What Is Blocking U.S. Power Expansion for AI—and What Could Unlock It by 2030?

Artificial intelligence (AI) is increasing demand for reliable electricity at a pace that is straining U.S. power infrastructure. Although many generation and grid projects are planned, bringing additional capacity online fast enough is difficult because of permitting delays, interconnection bottlenecks, transmission constraints, and limited flexibility from supplemental power sources. To understand what could realistically change by 2030, RAND researchers examined the barriers that most directly constrain the power grid’s net available capacity and identified the ones that federal policymakers and other stakeholders could plausibly address in the near term.

Mapping and Prioritizing Barriers to Power Expansion in the United States

The researchers began with 66 barriers drawn from a literature review and expert engagement. They then prioritized the barriers that could likely be addressed by 2030 and that the federal government could play a meaningful role in addressing. This process reduced the list to 17 priority barriers, which were grouped into four clusters of challenges, shown in Table 1.

For each cluster, the team developed a logic model (see the box below) to identify the stakeholders, actions, outputs, and outcomes needed to expand available power capacity. This process is summarized in Figure 1.

1. 66: Barriers collected from the literature
2. 17: Prioritization rubric based on timeline, agent, impact
3. 4: Grouping into common overarching themes

The first cluster is permitting delays. Large energy projects often require approvals from multiple agencies, and reviews can take years. Delays can arise from fragmented permitting responsibilities, lengthy environmental reviews, and land-access constraints. The report suggests that faster progress could come from better federal planning, the use of federal land for priority energy sites, and more-streamlined programmatic environmental reviews. These reforms could help bring projects online sooner and may unlock about 16 to 54 GW of additional capacity.

The second cluster is inefficient and costly interconnection processes, which the report identifies as one of the most important near-term bottlenecks. Interconnection studies are often slow, manual, and expensive, while existing interconnection rights are not always used efficiently. The report highlights opportunities to standardize generator replacement rules, improve transparency about underused interconnection capacity, and expand the use of surplus interconnection so that new resources can connect at existing sites without waiting for the full conventional process. Among the solutions examined, this cluster appears to offer the largest potential payoff, with estimated gains of roughly 65 to 130 GW.

The third cluster is underutilized transmission capacity. Existing transmission lines are not always operated as efficiently as possible, and grid operators may lack incentives to adopt technologies that improve performance without requiring major new construction. The report points to grid-enhancing technologies—such as dynamic line ratings, topology optimization, and power flow control—as a way to unlock more value from the current system. Although estimated gains are smaller and more uncertain than for some other actions, these measures could still improve system efficiency and speed up interconnections. The report suggests gains of up to about 7 GW, along with broader operational benefits.

The fourth cluster is the lack of options and incentives for supplemental generation. Data centers and other large loads often maintain backup generation, but existing rules and market incentives limit the extent to which these resources can support the grid during emergencies or periods of system stress. The report suggests that reforms to backup-generation regulations, support for cleaner nondiesel technologies, and tariff changes that encourage large loads to shift onto on-site or colocated generation during emergencies could create additional flexibility. Depending on implementation and participation, these measures could yield roughly 11 to 106 GW of effective additional capacity.

For each barrier cluster, the researchers used a logic model to analyze and illustrate the causes and courses of actions. For the second cluster, which pertains to challenges in interconnection processes, the logic model focuses on how to make better use of existing and underused interconnection capacity to bring new resources online more quickly. It centers on three actions:

• standardize generator replacement rules
• improve transparency and data on underused interconnection capacity
• create faster pathways for low-risk surplus interconnection projects.

These actions could reduce delays, lower costs, and help developers and grid operators make more-efficient use of existing grid infrastructure by 2030. The chart below illustrates the application of the logic model.

Problems or Barriers

• Lengthy, complex, and costly interconnection processes are a major bottleneck for bringing new energy resources online.
• Existing interconnection rights are underutilized because of inconsistent implementation, a lack of standardization, and insufficient market incentives.
• Fast-track options without requirements for costly network upgrades—such as surplus interconnection service and generation replacement service—have the potential to expedite interconnection and increase capacity by 2030.

Enablers

• Clear federal and state policy guidance and processes
• Transparent capacity-availability data
• Data access improvements

Inputs

• Interconnection processes and services
• Workforce and industry capacity
• Federal and state laws
• Regulatory processes
• Time
• Staff
• Expertise
• Materials
• Transmission constraints on interconnection

Stakeholders

• Utilities
• ISOs, RTOs
• FERC
• State regulators
• Developers
• DOE
• National Laboratories

Actions

• Expand generator replacement eligibility.
• Harmonize material modification rules.
• Revoke transferability restrictions.
• Evaluate system impacts of adding storage to existing projects.
• Proactively identify projects eligible for surplus interconnection.
• Disseminate eligibility and guidance.
• Publish transmission capacity data.
• Implement eligibility criteria for fast-track options.

Outputs

• Uniform generator replacement rules across ISO regions
• Increased applications for surplus- and partial-capacity services
• Publicly available transmission capacity maps

Outcomes

• Interconnection processes improve, and understanding of transmission capacity effects increases completion rates.
• DOE engagement with utilities, uniform generator replacement rules, and publicly available transmission capacity maps increase completion rates.
• Interconnection requests and approvals are responsive to data.

Anticipated risks: Diverted resources from interconnection reform, grid reliability risk, unequal fast-track access.

NOTE: DOE = U.S. Department of Energy; FERC = Federal Energy Regulatory Commission; ISO = independent system operator; RTO = regional transmission organization.

Interconnected Barriers Require Coordinated Actions

A central finding of the report is that these barriers are interconnected. Addressing one barrier in isolation is unlikely to have the same effect as addressing multiple related barriers at once. For example, streamlined permitting will have limited payoff if interconnection remains backlogged, and interconnection reforms work better when supported by transmission optimization and clearer incentives for flexible resources. For this reason, the researchers frame their recommendations as parallel courses of action, not stand-alone fixes.

The researchers estimate that, taken together, these combined actions could unlock roughly 92 to 297 GW of additional capacity by 2030. These estimates are uncertain and should be treated as ranges rather than precise forecasts. Still, these estimates suggest that reforms to interconnection, permitting, transmission utilization, and supplemental generation could materially expand the amount of power available for AI data centers and other growing loads—if policymakers act quickly and in a coordinated way.

Policy Implications

Expanding U.S. power capacity by 2030 will require coordinated action across multiple bottlenecks rather than isolated reforms. The report identifies four priorities: improve federal planning and permitting for AI-related energy infrastructure, increase transparency and fast-track options in interconnection, incentivize adoption of grid-enhancing technologies, and support supplemental generation and demand response during emergencies. More specifically, it recommends

• improving federal planning and permitting by identifying suitable federal land for AI power sites, using programmatic environmental review where appropriate, and streamlining the permitting process for priority projects
• improving interconnection transparency and efficiency, including assessing and publicizing underused interconnection capacity, standardizing practices for surplus interconnection, and incorporating surplus interconnection opportunities into transmission planning
• encouraging adoption of grid-enhancing technologies through regulatory incentives, pilot programs, and collaboration among grid operators, utilities, and large loads
• reviewing barriers to supplemental and backup generation, supporting cleaner nondiesel technologies, and updating tariff structures so that large loads have stronger incentives to reduce grid demand or shift to on-site generation during emergencies.

Available for Download

Download PDF

Topics

• Alternative and Renewable Energy
• Artificial Intelligence
• Energy Consumption
• Energy Distribution
• Science and Technology