미국 정부가 국가 전력망 업그레이드를 위해 $5.25 billion의 지출을 발표했으며, 26개 주의 31개 프로젝트를 지원하여 새로운 AI 데이터 센터에 전력을 공급한다.
The US Department of Energy announced on September 24 that it intends to fund 31 transmission projects under its Speed to Power through Accelerated Reconductoring and other Key Advanced Transmission Technology Upgrades (SPARK) initiative. The $5.25 billion project value represents a combined effort: the DOE is committing $1.9 billion in grants, roughly 36% of the total, with utilities, cooperatives, state agencies, and selected companies providing the remaining $3.35 billion.
The work will rebuild more than 1,500 miles of transmission line, add grid-enhancing technologies to nearly 21,000 additional miles, and free up over 23 GW of transmission capacity. Energy Secretary Chris Wright characterized the investments as a means to "get more out of the infrastructure we already have" as the country contends with an aging grid largely built 50 to 75 years ago. The DOE expects to finalize the awards between October 2026 and January 2027, though it reserves the right to redirect or discontinue funding at go/no-go checkpoints.
SPARK is the third round of the Grid Resilience and Innovation Partnerships (GRIP) program authorized by the 2021 Infrastructure Investment and Jobs Act. The DOE's official announcement does not mention artificial intelligence or data centers, yet the initiative is explicitly tied to AI infrastructure priorities: the agency launched the wider Speed to Power program in September 2025 partly to "win the global artificial intelligence race," lists SPARK among its data center grid actions, and counts enabling data centers as a qualifying benefit.
The connection between this funding and AI demand, however, remains largely indirect. Only three of the thirty-one project summaries name AI or data centers as direct beneficiaries, representing roughly 7% of the total project value, or $371 million. A larger portion of the funding—roughly half the total—sits in four state-led projects classified under the category of "transmission systems that facilitate development of new large loads," an industry term encompassing data centers, factories, and other heavy power consumers.
Adding large numbers of AI data centers could quickly overwhelm the grid at various junctions, making power delivery and transmission capacity as critical as raw compute capacity in the US competition against Chinese advances. With 23 GW of transmission capacity being unlocked nationwide, and private and public generation being scaled up to meet AI power demand, energy costs are likely to rise. This occurs at a time when crude oil hovers near $100 per barrel due to prolonged Middle East conflict, and utility companies are positioning infrastructure investments that will eventually appear on residential power bills. While federal funding may moderate these increases, small domestic consumers and taxpayers will likely absorb much of the cost for a buildout that does not appear to involve data center developers directly financing the capacity they will ultimately consume.