FERC Orders Grid Operators to Speed Up Data Center Power Hookups

The growth of artificial intelligence is consuming electricity at a pace that electrical grids have not seen before. Data centers that train and run large AI systems now compete aggressively for limited power supply in the regions where they cluster, forcing regulators and utilities to make rapid decisions about who gets access to new capacity.

For technology companies planning infrastructure expansion, grid availability is becoming as important as land, real estate costs, and water access. States and grid operators are implementing new rules to manage demand, shifting infrastructure costs to developers and delaying deployments by months or years. The Federal Energy Regulatory Commission stepped in during June 2026 with orders requiring major grid operators to reform their interconnection processes or justify why they cannot accommodate large data center loads.

The scale of electricity demand

Data centers consumed approximately 4.7 percent of total U.S. electricity consumption in 2024, or 192 terawatt-hours, according to a 2026 Department of Energy study cited by the Congressional Research Service. The study projects that by 2030, data center consumption will reach between 521 and 843 terawatt-hours annually—representing 9.5 to 15.3 percent of U.S. total consumption.

How grid connection actually works

Connecting a data center to the electrical grid is a multi-step process managed by regional grid operators, not utilities. A developer begins by entering the interconnection queue—essentially a backlog of projects waiting for technical study. The grid operator must conduct an interconnection study to determine whether the new data center would pose reliability risks to the system. Since 2003, the Federal Energy Regulatory Commission has set requirements for these studies, establishing timelines and standards that were designed for power plants, not massive computational loads.

The process takes significantly longer than data center construction. Research from Marcus by Goldman Sachs found that only 72 percent of data centers scheduled for activation within four quarters historically go online on time, largely due to grid connection bottlenecks.

Regional grids at capacity limits

Fifteen states account for 80 percent of national data center electricity use, according to the Congressional Research Service, creating acute pressure on specific regional grids. Virginia alone consumes over 20 percent of its state's electricity for data centers.

Texas faces the most visible crisis. The state's interconnection queue—the backlog of power plants and infrastructure projects waiting to connect to the grid—contains 474 gigawatts of new capacity requests, according to reports from September 2026. Data centers comprise roughly 90 percent of these applications. Critically, many of these requests are speculative duplicates. Developers frequently file duplicate interconnection applications across multiple regions to hedge against delays, according to research from Marcus by Goldman Sachs. Amid concerns about this practice, the state's governor ordered a pause on data center development in August 2026 pending an audit of the queue.

New York implemented the first statewide moratorium on new data center development in July 2026, temporarily pausing state environmental permits for up to a year to build what officials called "a nation-leading regulatory framework that protects ratepayers, the environment, the energy grid and communities across the state." PJM Interconnection, which operates the grid serving Virginia and the mid-Atlantic region, proposed requiring large power users to contract for their own generation capacity or face curtailment during shortages—a shift that places infrastructure costs directly on data center developers rather than spreading them across all utility customers.

Federal intervention to accelerate interconnection

In June 2026, the Federal Energy Regulatory Commission issued six "show cause" orders requiring regional transmission organizations and independent system operators—PJM, MISO, CAISO, SPP, NYISO, and ISO-NE—to address inadequacies in their large load interconnection rules. FERC Chairman Laura Swett stated the agency acted to promote "fair cost allocation, unprecedented transparency for the American ratepayer, respect for states' rights, efficient markets and speed to power."

FERC directed grid operators to address five specific issues within 60 days: streamline transmission service application processes and evaluate alternative transmission technologies; prevent cost-shifting by clearly accounting for network upgrade expenses; develop interconnection options for co-located or behind-the-meter generation; create transmission services for data centers that can curtail operations during peak demand; and establish faster study processes for facilities serving electrically proximate loads. Grid operators could request 90-day extensions, but the agency made clear that failure to comply could result in FERC dictating solutions unilaterally. The June order represents an escalation from preliminary reforms FERC had implemented in 2023, which introduced "readiness requirements" designed to reduce speculative interconnection requests and cluster studies to evaluate multiple projects together.

Cost and pricing impacts across the region

Historical analyses from 2019 through 2025 showed that states with the largest data center growth generally saw electricity price decreases, according to Lawrence Berkeley National Laboratory research cited by the Congressional Research Service. One explanation is that increased demand allowed utility costs to be spread over a larger sales volume, putting downward pressure on rates.

Future modeling, however, predicts very different outcomes. A modeling study led by North Carolina State University, cited by the Congressional Research Service, found that data centers could increase national average electricity costs by 6 to 29 percent across different modeling scenarios, with increases up to 57 percent in some regions.

How major tech companies are responding

Technology companies are attempting to sidestep grid constraints by securing long-term power supplies independently. Microsoft signed a $10 billion renewable energy deal with Brookfield Asset Management deploying over 10.5 gigawatts of capacity starting in 2026, and entered into a purchase agreement with fusion energy developer Helion Energy for electricity delivery by 2028. Amazon purchased a nuclear-powered data center campus in Pennsylvania for $650 million, while also advancing liquid cooling technologies that reduce mechanical energy consumption. Alphabet (Google) signed contracts for 8 gigawatts of clean energy and partnered with Kairos Power for small modular reactors. Meta is developing a $10 billion AI-focused data center in Louisiana combining renewable energy with next-generation nuclear technology.

These strategies serve dual purposes: they bypass congested grid interconnection queues by pairing data center development with dedicated generation, and they allow companies to maintain renewable energy commitments while scaling operations.

Grid reliability and frequency stability

Data centers do not appear to have caused blackouts to date, but their size creates new risks during grid faults. The North American Electric Reliability Corporation identified in 2026 that sudden disconnections of large computational loads pose stability threats. More than 1,000 megawatts of data center load unexpectedly disconnected during transmission faults in recent incidents, creating frequency disturbances that threaten grid integrity.

The risk arises because data centers respond instantly to power disruptions by tripping offline automatically to protect equipment, failing to ride through transmission disturbances that are normally cleared without incident. This instant load loss can destabilize the frequency of the grid itself, potentially cascading into problems for other users.