• The US Department of Energy has plans for $1.9 billion in grants for 31 grid projects worth $5.25 billion, with recipients covering the other $3.35 billion
• The money is essentially rebranded Grid Resilience and Innovation Partnerships (GRIP) funding from the 2021 infrastructure law, aimed at reconductoring and grid-enhancing tech rather than a focus on new power plants
• Only three project summaries tie the work to AI or data center demand, but half the value of the spending sits in four state-led projects under a category DOE designed around ‘new large loads’, which could potentially be addressing AI data center demand

The US Department of Energy’s Office of Electricity said on September 24 that it intends to help fund 31 transmission projects under its Speed to Power through Accelerated Reconductoring and other Key Advanced Transmission Technology Upgrades initiative, a mouthful that it has somewhat mercifully shortened to SPARK.

The $5.25 billion attached to the announcement, however, is the projects’ combined value, not federal spending, with the federal government only picking up roughly 36% of the tab.

The DOE’s $1.9 billion is tied to commitments from utilities, cooperatives, state agencies, and companies it selected that will provide the remaining $3.35 billion in coverage.

An AI-centric play?

The DOE’s announcement focuses on the infrastructure side of the equation; it says the work will rebuild more than 1,500 miles of transmission line, add grid-enhancing technologies to nearly 21,000 more, and free up over 23 GW of capacity.

The Energy Secretary Chris Wright said the investments will “get more out of the infrastructure we already have,” even as the current documentation lists selections rather than actual signed awards.

The DOE expects to finalize them between October 2026 and January 2027, according to its SPARK program page, and the funding notice lets the department redirect or discontinue funding at go/no-go checkpoints.

The money isn’t exactly new, nor is the idea for a country dealing with an aging grid that saw most of its infrastructure built 50 to 75 years ago; SPARK is the third round of the Grid Resilience and Innovation Partnerships (GRIP) program funded by the 2021 Infrastructure Investment and Jobs Act.

Adding a large number of AI data centers would quickly overwhelm it at various junctions, potentially turning the AI infrastructure race into one about power delivery and capacity versus actual compute power as the US continues to fend off Chinese advances in the arena.

The DOE’s release never mentions artificial intelligence or data centers, though publications like The Register were less subtle, saying “Bit barn builders can rejoice” as a result of the new commitment. The link is indirect but present. DOE launched the wider Speed to Power initiative in September 2025 partly to “win the global artificial intelligence (AI) race”.

Its data center resource hub lists SPARK among its grid actions, and the funding notice counts enabling data centers as a qualifying benefit. This is despite only three of the thirty-one project summaries naming AI or data centers outright as direct beneficiaries. They amount to roughly 7% or $371 million of the total funding in play here.

Bigger projects, however, are more discrete but leave no doubt about where the money is going; they focus on building “transmission systems that facilitate development of new large loads,” the industry’s catch-all for data centers, factories, and other heavy users.

With potentially 23 GW of transmission capacity essentially being unlocked on the grid, coupled with private and public generation being geared up across the country to meet AI power demand, one could assume- and correctly so that energy costs will only increase at a time when crude oil is hovering near the $100 mark due to a prolonged conflict in the Middle East, coupled with large spending being earmarked by utility companies that will inevitably show up on residential power bills.

Federal funding for the grid might dull the blow, but small domestic consumers and taxpayers will inevitably pick up the tab for a buildout that doesn’t seem to have data center developers lining up to pay for the capacity they will eventually consume.