Data Center Power Supply Gap Puts Gas Pipelines and Utilities in Focus

The AI buildout is running into a physical limit that capital spending alone cannot solve, and that limit is electricity. Data center power availability is increasingly constrained in parts of the U.S., so the builders that have already secured energy through a natural gas pipeline or a traditional utility hold the advantage in bringing AI capacity online. That moves gas pipelines, utilities and the U.S. grid into the same conversation as chips and cloud spending, just as political opposition to data centers adds scrutiny to every new project.

Yi Fu Lee, StoneX Financial Inc. Managing Director and Senior Equity Research Analyst, covers AI, cybersecurity, infrastructure and enterprise software and technology. That coverage gives insight into how energy access, grid capacity and funding are shaping the AI data center buildout.

Key Themes from the Discussion

• AI data centers need land, a building, electricity and chips, and electricity is the input in shortest supply.
• Builders with secured natural gas pipeline or utility supply could earn a valuation premium for their data centers.
• Consumers, ratepayers, utilities and data center builders all share the cost of the grid upgrades AI requires.

Watch the Full Conversation

Data Center Power Supply Determines How Fast AI Capacity Comes Online

Data center power supply is the gating input for AI, because a facility with land and chips cannot run a single workload without electricity behind it. "You need the land itself, the data center itself, the power, the electricity and the chips," Lee explains, likening an AI data center without power to a phone without a battery. Research from the U.S. Department of Energy that Lee cites points to 104 gigawatts of firm generation slated to retire by 2030, while DOE warns that planned replacement supply is insufficient as electricity demand accelerates. As a result, the pace of the AI buildout depends as much on access to electricity as on chips or capital.

Secured Energy Gives Data Center Builders a Valuation Edge

Data center builders that have locked in their energy supply hold a competitive edge while U.S. power capacity runs short. According to Lee, securing that energy is "extremely pivotal, and you could probably earn a premium in terms of your valuation" for both the data center and the company behind it. Specifically, the advantage sits with builders that have already arranged supply through a natural gas pipeline or a traditional utility connection. For investors, energy access becomes part of how AI infrastructure companies are valued, alongside chips, land and capital.

A builder with secured supply is holding "that card in (their) hand," in Lee's words, with a natural gas pipeline or utility supply already behind the project. Builders without that card face a slower route, since policymakers can place additional restrictions on permitting and energy allocation to protect local communities. Lee does not see those limits as a bad thing, arguing that lawmakers and regulators should act to protect their own citizens. Consequently, new data-center providers may face more steps in the approval process, which raises the value of energy supply a builder has already secured.

U.S. Grid Upgrade Costs Fall on Consumers, Ratepayers and Utilities

The cost of expanding U.S. data center power supply falls on consumers, ratepayers, utilities and data center builders together. Ultimately, Lee says, "the brunt of it is going to be the consumers," with developers, ratepayers and utility companies also sharing the bill. Policymakers shape how fast that spending happens, because lawmakers and regulators can limit licenses and energy allocations to protect local communities. Lee treats that cost as one society carries collectively. "There's no free lunch, and in this case, no free dinner as well."

Lee looks at the AI buildout from a glass half full perspective as it is a chance for America to modernize its core infrastructure. "Not only are we upgrading our AI infrastructure, but we should also upgrade our energy utility infrastructure. This is our chance to do it." Lee argues the U.S. grid needs an upgrade, and the scale of AI capital spending creates an opening to fund one alongside new data centers. Conversely, slower approvals and tighter licensing determine where that investment lands and how quickly it arrives. For utilities and gas pipeline operators, the AI buildout therefore brings both new electricity demand and pressure to modernize the grid.

--- Written by Mohamed Faisal, StoneX Media Editor

--- Expert: Yi Fu Lee, StoneX Financial Inc. Senior Research Analyst