Who Controls Data Center Power?
Understanding the Invisible System Behind Government Infrastructure
As demand for data centers accelerates—driven by AI, cloud expansion, and government modernization—one constraint is rising above all others:
Power.
Not just how much is needed, but who controls it, how it is allocated, and how quickly it can be delivered.
What’s becoming clear is that there is no single authority making these decisions. Instead, power for data centers is governed by a layered, often fragmented system. One that was never designed for the scale and speed of today’s demand.
To understand where the friction is coming from, you have to understand how the system actually works.
At the top sits the Federal Energy Regulatory Commission, or FERC. It doesn’t approve data centers or decide where they get built. Instead, it sets the rules that govern the flow of electricity across state lines and defines how wholesale energy markets operate.
These rules determine how transmission is planned, how pricing works, and critically, how large new loads connect to the grid. In effect, FERC shapes the economic and regulatory environment in which data centers either move forward or stall.
But the real friction begins one layer below.
Regional grid operators, organizations like PJM Interconnection, ERCOT, and CAISO, are responsible for keeping the grid balanced in real time. They manage supply and demand, operate electricity markets, and oversee the interconnection process for new projects.
This is where theory meets reality.
As data center demand surges, these operators are facing unprecedented volumes of large-load interconnection requests. The result is a growing backlog—projects waiting years, not months, for approval. In many cases, new data centers are being told that the grid simply cannot accommodate them without significant upgrades.
Even when approval is granted, the next challenge emerges: actually delivering power to the site.
That responsibility falls to utilities, the entities that build and operate the infrastructure that physically connects a data center to the grid. Utilities determine how quickly power can be delivered, what infrastructure needs to be built, and how much it will cost.
For large-scale data centers, this often means new substations, transmission extensions, or long-term capacity commitments. In practice, utilities have become the de facto gatekeepers of growth, the ones who ultimately decide whether a project can move forward on a realistic timeline.
But utilities don’t act alone.
They are regulated at the state level by public utility commissions, which approve rates, oversee infrastructure investments, and increasingly, weigh in on whether large energy users are placing undue burden on the system. As data center demand grows, these regulators are asking harder questions: Who pays for grid upgrades? Should residential customers subsidize large-scale compute infrastructure? And what is the long-term impact on reliability?
These questions are beginning to reshape policy in real time.
And then, at the final layer, there are local governments—counties and municipalities that control zoning, permitting, and community approvals. Even when power is available, projects can be delayed or blocked due to land use concerns, environmental reviews, or public opposition.
Taken together, this system reveals a fundamental truth:
Data center power is not controlled—it is negotiated.
Across multiple layers. Across multiple timelines. And increasingly, across competing priorities.
And that system is now under strain.
Interconnection queues are growing. Transmission infrastructure is lagging. Policy frameworks are evolving. And AI-driven demand is accelerating faster than any of these systems were designed to handle.
For government data centers, this has immediate and long-term implications.
Power is no longer just an input. It is the primary constraint. Project timelines are no longer predictable, they are subject to structural delays. And traditional approaches to siting and development are no longer sufficient.
In response, a new set of strategies is beginning to emerge. Agencies and operators are exploring behind-the-meter energy solutions, on-site generation and storage, and more distributed architectures that reduce dependence on any single point in the system. Others are rethinking location strategy entirely, prioritizing regions where power availability aligns with mission requirements.
What’s changing is not just where data centers are built, but how they are conceived from the outset.
Gov DCx POV
The conversation around data centers has shifted.
It is no longer about compute capacity, cloud strategy, or even physical infrastructure.
It is about power; how it is governed, how it is delivered, and who ultimately controls access to it.
And in today’s environment, the answer is not simple.
The organizations that succeed will not be those that simply build faster. They will be the ones that understand the system, navigate it effectively, and design around its constraints.
Because in the next phase of digital infrastructure:
Power is not just a dependency. It is the strategy.
