DOE Unveils $1.9B Funding to Expand Grid Capacity for AI
US SPARK initiative funds 31 transmission projects to unlock 23 GW of power
The U.S. Department of Energy (DOE) has awarded $1.9 billion in federal funding across 31 grid-improvement projects to accelerate transmission capacity for power-hungry AI data centers. Part of the new SPARK initiative, the combined $5.25 billion public-private investment aims to unlock over 23 gigawatts of additional electrical grid capacity across 26 states.
Key details
The funding, administered by the DOE's Office of Electricity under the Speed to Power through Accelerated Reconductoring and other Key Advanced Transmission Technology Upgrades (SPARK) initiative, pairs $1.9 billion in federal grants with $3.35 billion in recipient cost-share capital. The 31 selected projects focus on maximizing the throughput of existing transmission infrastructure without waiting years for new long-distance lines to be built.
Project plans include reconductoring or rebuilding more than 1,500 miles (2,412 kilometers) of high-voltage transmission lines with high-capacity advanced conductors. Additionally, grid-enhancing technologies (GETs) and Advanced Transmission Technologies (ATTs) will be deployed across nearly 21,000 miles (3,379 kilometers) of lines, enabling real-time dynamic line rating and power flow control.
Key utility projects explicitly targeting data center demand include Alabama Power's Northeast Alabama Smart Grid Deployment, which received $50 million to upgrade transmission lines supporting regional data center expansion. Similarly, Entergy Mississippi secured funding for a three-state smart-grid deployment across Arkansas, Louisiana, and Mississippi designed to release latent grid capacity for hyperscale industrial loads.
Why this matters
Grid congestion and multi-year interconnection queues have become the primary bottleneck for expanding AI compute infrastructure in North America. By focusing on reconductoring and GETs rather than greenfield transmission builds, the DOE initiative provides a faster path to power for gigawatt-scale data center campuses. Unlocking 23 GW of existing grid capacity helps relieve regional power bottlenecks that have increasingly threatened localized grid stability and driven up wholesale electricity prices.
Context
As hyperscalers and neocloud providers deploy massive AI clusters, power demand has outpaced utility forecasting across major regions like PJM, ERCOT, and the Southeast. Prior federal efforts centered on long-term transmission planning, but the rapid surge in AI electricity consumption has forced energy regulators and utilities toward immediate efficiency gains on existing lines.
Risks and open questions
While advanced conductors and GETs offer rapid capacity gains, questions remain regarding how quickly regional transmission operators can integrate these technologies without causing operational disruptions. Furthermore, local communities and consumer advocates continue to watch closely to ensure that public funding and utility cost-shares do not shield commercial data center developers from paying their fair share of underlying grid upgrades.
What happens next
Recipients will finalize cost-share agreements and begin procurement for advanced conductors, sensors, and control systems. Utility operators plan to start physical line rebuilds and GET deployments over the next 12 to 24 months, with initial grid capacity unlocks expected to take effect starting in late 2027.
Source: Data Center Dynamics Published on AI Usage Global, author: AUG Bot
