There is a massive shift underway in how we generate and use electricity. More renewables are coming online every day, and one-way grids are becoming more like networks with multi-directional power flows from an ever-increasing number of resources.
On the other side of the equation, it’s all about growth, not only from the exponential increase in demand from data centers but also from new sectors like EV charging and the ongoing electrification of industrial processes. The result is skyrocketing demand growth over a brief period of time, which is unlike anything the power industry has seen since its beginnings in the 19th century.
So, how do we meet the challenge?
It starts with efficiency. There are solutions on the supply side, but we should first remember the industry adage that “the least expensive megawatt is the one you don’t have to generate.” As a leading energy technology provider, Schneider Electric has a long history of developing and deploying innovative solutions to increase energy efficiency and optimization across industrial, commercial, and residential environments.
Let’s take a look at how this affects data centers.
The data center dilemma
A recent Accenture report forecasts that data centers could consume over 7% of total U.S. electricity by 2028, rising sharply to 16–23% by 2033. This marks a substantial jump from their 5% consumption in 2024, highlighting the growing energy demands of digital infrastructure. The growth is uneven, with regions like Northern Virginia, Texas, and California nearing the limits of their local grids, while others remain underutilized. Adding to the complexity, AI workloads, often referred to as “AI factories,” consume significantly more power than traditional compute environments, creating high-density demand that strains local power infrastructure.
Building new generation capacity takes years, and new transmission infrastructure takes even longer. Some data centers face wait times of four years or more just to connect to the grid. Meanwhile, AI adoption is accelerating, creating a feedback loop of rising demand and constrained supply.
Efficiency is for everyone
Reducing demand is faster, more efficient, and less expensive than building new generation capacity (though that is also needed). In industrial environments, efficiency improvements can significantly lower energy consumption and associated costs without compromising performance.
For example, consider the systems that distribute power throughout a factory or process facility. Minimizing losses and risks while ensuring optimized performance and reliability are top priorities. Connecting smart technologies like automatic transfer switches (ATS), uninterruptible power supplies (UPS), and intelligent switchgear and cooling equipment (e.g., chillers, variable speed drives) to advanced digital software platforms makes power and cooling systems more robust and more efficient.
In data centers, cooling is a challenge that grows as servers become more electrically intensive and powerful. IT rack densities driven by more powerful GPU architectures are now forecasted with systems designed for 600KW – 1MW per rack. That’s why we invested in Motivair, the leader in liquid cooling for high-density data centers. Now, more than ever, we’re actively engaged with firms to design optimized cooling strategies for new chip designs before they hit the market.
We’re leveraging AI itself, too, with tools that analyze energy usage patterns, identify inefficiencies, and automate corrective actions in all kinds of facilities. These systems can detect anomalies, predict failures, and optimize operations in real time.
One proof point is Compass Data Centers’ use of analytic tools from Schneider and AVEVA to enable predictive analytics and condition-based maintenance on electrical system assets. The results are compelling: a 40% reduction in on-site maintenance interventions and 20% lower operational costs.
We are also working with Nvidia to develop standardized, scalable reference designs for data centers that can be easily built and repeated at various locations. The designs are publicly available and will help accelerate the deployment of new data centers while ensuring they are as energy efficient as possible.
Digitalization is the foundation
Whether in manufacturing, process industries, or any other sector, the advances in AI-ready infrastructure we’re bringing to market are all predicated on digitalization. AI (and analytics more broadly) relies on data, and the more that’s available from operational, monitoring, and control systems, the better.
For instance, digital twins are virtual copies of facilities created using operational data. By applying predictive analytics, operators can simulate different scenarios on the model, forecast demand, and identify opportunities to eliminate waste. These tools are essential for proactive energy management (i.e., greater efficiency), but they also boost productivity and cut downtime.
As for data centers, applying currently available technologies can reduce overall facility demand by as much as 17%.
The business case for a combination of efficiency and digitalization is strong, and Schneider is uniquely positioned to guide customers along their digitalization journey toward a future that is more efficient and cost-effective.
These are just a few of the ways that Schneider is working to address the rise of AI in the new energy landscape. The challenges are real, but they’re solvable. By prioritizing efficiency, embracing innovative technologies, and fostering collaboration between industry players, North America can meet its fast-growing electricity demand while supporting sustainable growth.
Find out more about how Schneider is working to meet the challenge of data center demand.
Add a comment