How Big Is This?

The scale of data center deployment in the United States has moved from background infrastructure to headline national news in a matter of months. The numbers tell a story of explosive growth and unprecedented infrastructure demands.

4,200+
Data Centers Operating in the U.S.

Source: DataCenterMap.com

176 TWh
Electricity Consumed in 2023 (4.4% of U.S. Total)

Source: DOE/LBNL, Dec 2024

325–580 TWh
Projected Demand by 2028

Source: DOE/LBNL, Dec 2024

The Trajectory

In just five years, data center electricity demand is projected to nearly triple. The following chart shows the current state versus the low and high projections for 2028:

2023 Actual
176 TWh
176 TWh
2028 Low
325 TWh
325 TWh
2028 High
580 TWh
580 TWh

What This Means

Even the conservative 2028 projection represents an 85% increase in just five years. At the high end, data center electricity demand could nearly double from 2023 levels. This surge is unprecedented in the recent history of the U.S. electrical grid.

The Grid Needed Investment. Data Center Demand Is Forcing It to Happen.

For the better part of two decades, the U.S. electrical grid operated under a comforting assumption: electricity demand would remain relatively flat. Planning was conservative. Infrastructure investment was measured. Meanwhile, the grid itself aged. That assumption is now broken—and the result is forcing necessary modernization, though at a pace that raises real questions about fairness and cost allocation.

A Pattern That Held for Nearly Two Decades

From roughly 2005 to 2023, U.S. electricity demand was essentially stagnant. Efficiency gains, demand-side management, and economic shifts kept consumption level despite population growth. The grid had time to adapt. Utilities planned for stability, not growth.

Data centers are the first major demand sector to break this pattern. Unlike previous waves of electricity-intensive industry, data centers are growing exponentially, require enormous amounts of power in concentrated locations, and are not geographically bound by traditional industrial constraints.

The Numbers Behind the Crunch

The U.S. has approximately 1,300 GW of total installed electrical capacity. Data centers are projected to add 50 GW of demand by 2030 alone — nearly 4% growth driven by a single sector. The grid has never had to absorb growth of this magnitude from one industry in such a short timeframe.

Source: EPRI

The Capacity Crisis

In regions like Pennsylvania, PPL Electric projects a 60 GW demand and only 54 GW of capacity by 2030 — a 6 GW shortfall. This is not theoretical. Utilities are already warning that without new generation and transmission, they cannot meet demand from data centers that are already under construction.

This is why the power plants being built in St. Charles Parish matter. They are not just corporate investment. They are part of the national effort to modernize a grid that has been neglected for two decades and is now facing genuine constraints. The speed is being driven by data centers, but the underlying need for investment is real.

AI Is the Engine

Behind the data center explosion is a single technology: artificial intelligence, and specifically the race to build, train, and deploy large language models and generative AI systems at massive scale.

The Power Requirements of AI

Training a single large language model consumes extraordinary amounts of electricity. OpenAI's GPT-4, for example, required approximately 25,000 megawatt-hours of electricity for the training phase alone. That is enough to power 2,000 typical American homes for an entire year — just to train one model.

But training is only the beginning. Once deployed, AI inference — the process of running trained models to answer millions of user queries every second — runs continuously, 24 hours a day, 7 days a week. A single large-scale AI data center dedicated to inference can consume 300+ megawatts continuously.

25,000
MWh to Train GPT-4 (One Model)

Source: OpenAI

300+ MW
Continuous Power for Large AI Data Center

Source: Industry reports

$2.6–4.4T
Estimated Economic Value of Gen AI

Source: McKinsey

The Economic Prize

McKinsey estimates that generative AI could unlock $2.6 to $4.4 trillion in economic value. To realize this potential, industry experts project that 50 to 60 GW of new data center capacity will be needed over the next several years. This infrastructure does not yet exist. It must be built.

Why This Matters to St. Charles Parish

The power plants being built in St. Charles Parish are directly tied to this AI infrastructure race. These are not traditional industrial facilities. They are purpose-built to support the generation, training, and deployment of AI systems at a scale that will define the next decade of technology and the economy.

Water: The Hidden Cost

As critical as electricity is to data centers, another resource is equally vital: water. Data centers consume massive amounts of water for cooling, and the scale of this consumption has only recently become visible in public discussions.

Direct and Indirect Consumption

Data centers consume water in two ways:

  • Direct cooling: Water is cycled through data center equipment to dissipate heat, then discharged or recycled.
  • Indirect consumption: The electricity used by data centers must be generated somewhere, and most power plants require vast amounts of water for cooling.
17B
Gallons of Water Consumed by Data Centers in 2023

Source: EESI, World Resources Institute

34B+
Gallons Projected by 2028 (Double)

Source: EESI, World Resources Institute

211B
Gallons Indirect (Generation & Cooling)

Source: EESI

A Decade-Scale Impact

By the end of the decade, Yale E360 reports that U.S. data centers could consume as much water as 10 million Americans combined. In arid and water-stressed regions, this is already driving conflict and regulation.

Louisiana's Water Advantage — and Hidden Risk

Louisiana has abundant water. It sits at the mouth of the Mississippi River. This is one reason why data center operators have targeted the state. But Louisiana has a critical blind spot:

No Mandatory Water Reporting for Data Centers

Unlike other states, Louisiana does not require data center operators to report water consumption. This means there is no public accounting for how much water these facilities actually use. Meta's St. Charles Parish facility is registered to consume 8.4 billion gallons annually, but the company's own reports suggest actual annual usage is 500–600 million gallons. Without transparency, residents and policymakers cannot assess the true impact on local water resources.

The Transparency Question

Other states have begun to mandate data center water disclosure. California, for example, requires detailed reporting. Louisiana does not. As more data centers come online in St. Charles Parish and beyond, the lack of mandatory reporting becomes a growing risk to water security and environmental stewardship.

Why Louisiana?

The U.S. data center explosion could have targeted any state. Instead, more than $41.5 billion in data center investment has been directed to Louisiana in recent years. This is not random. It is the result of a convergence of factors that make Louisiana uniquely attractive to industry — and uniquely vulnerable to their demands.

Other States Are Running Out of Capacity

In developed data center markets like Virginia, Pennsylvania, and Texas, grid capacity is already constrained. Utilities are warning that additional data center demand cannot be met without major new generation. This has driven operators to look elsewhere.

PPL Electric's projections are a bellwether: the Pennsylvania utility expects 60 GW of demand by 2030 but only 54 GW of capacity. Similar shortfalls are appearing across the Northeast, Midwest, and parts of Texas.

Louisiana Has What Others Lack

Louisiana offers a rare combination of assets:

Available Land and Infrastructure

Louisiana has industrial sites and existing power plants that can be repurposed or expanded. The state has the infrastructure skeleton to support rapid data center deployment.

Abundant, Inexpensive Water

Access to the Mississippi River and aquifers provides water for both direct cooling and power generation. This is a competitive advantage that few states can match.

Willing Regulators

Louisiana has, for decades, prioritized industrial development. Regulatory frameworks are streamlined for projects that create jobs and tax revenue. This extends to data center permitting and land use.

Cheap Power

Louisiana's electricity rates are among the lowest in the nation, driven by legacy natural gas infrastructure and hydroelectric access via the Southwest Power Pool. Cheap power is a direct incentive for data center operators, whose operating costs are dominated by electricity.

St. Charles Parish as Ground Zero

St. Charles Parish has all of these advantages. It is near existing industrial infrastructure, has access to water, and is positioned within a region that can generate and distribute significant power capacity. $41.5 billion in investment has flowed to Louisiana, and a significant share is landing in St. Charles Parish because the parish is well-positioned to deliver on the infrastructure needs that the nation now urgently requires.

The Convergence

The demand explosion is not slowing. AI continues to grow. The grid is constrained and aging. Louisiana has capacity and willingness to invest. The power plants being built in St. Charles Parish are not an anomaly. They are the inevitable result of national infrastructure scarcity meeting regional advantage. The modernization is largely necessary and good — but the speed and who bears the costs are the real policy questions.