We've spent years highlighting the transformative power of data. Here at Gorilla, it's the bedrock of everything we do. So, when we see headlines about the explosive growth of data centers – the very infrastructure housing this data – our first instinct might be to cheer. But a shadow hangs over this boom. The primary engine driving this expansion isn't just the general need for data storage; it's the specific, colossal computing requirements of Artificial Intelligence, especially generative AI. This AI gold rush comes with a hefty price tag: an almost insatiable demand for electricity. This escalating energy consumption poses a significant threat, potentially overwhelming national power grids.

Data centers across the United States are exhibiting a significant increase. In 2023, these facilities accounted for approximately 4.4% of total US electricity usage, having more than doubled their demand since 2017.  Projections indicate this trend will continue sharply upwards; estimates suggest data centers could consume between 6.7% and 12% of US electricity by 2028, potentially reaching nearly 12% or even higher by 2030.

Within these digital pages, we’ll shed light on this shift, why it matters for energy retailers in the ERCOT region and beyond, and how to plan for a data-hungry future. 

Data shoots past competition

For the past decade, much of the conversation around new electricity demand has focused on the electrification of transport, specifically the rise of electric vehicles (EVs). While EVs continue to represent a significant and growing load, the landscape is rapidly shifting. The exponential growth in cloud computing and the recent explosion in AI has propelled data centers to the forefront of new energy demand.

• The Scale of the Shift: Recent industry analyses and surveys of utility executives highlight this dramatic change. For instance, some studies indicate that utility leaders now attribute a larger portion of anticipated load growth directly to data centers and AI – potentially around 40% - compared to the combined growth expected from sustainability initiatives like EV charging and residential solar (around 31%). This signifies a fundamental change in demand drivers.
• Why the Surge? AI model training and inference are incredibly energy-intensive processes, requiring vast amounts of computational power housed in specialized data centers. Simultaneously, the ongoing migration of business operations to the cloud and the increasing reliance on digital services continue to fuel demand for traditional data center capacity.
• The New "Electric Load on the Block": This trend firmly establishes data centers, powered by countless servers processing complex algorithms and storing vast datasets, as the primary new large-scale consumers of electricity in many regions. The energy footprint of the digital world is rapidly outpacing the electrification of roads, presenting both immense challenges and significant opportunities for energy retailers trying to forecast demand and manage grid resources.

Geographic hot spots and grid strain

The surge in data center demand isn't evenly distributed across the grid; instead, it's highly concentrated in specific geographic "hot spots." These clusters create intense, localized strain on the electricity infrastructure.

• Key Hubs:
  
  • Northern Virginia ("Data Center Alley"): Historically the world's largest concentration of data centers due to its proximity to Washington D.C. and foundational fiber optic networks. It continues to see massive growth, pushing regional utilities like Dominion Energy to navigate unprecedented demand requests and transmission constraints. Reports have surfaced about temporary pauses or delays in connecting new, large data centers due to insufficient transmission capacity in certain areas.
  • ERCOT: The state's business-friendly environment, availability of land, and significant (though sometimes intermittent) renewable energy resources have made it attractive. Texas is seeing a boom not only in traditional cloud and AI data centers but also in energy-hungry cryptocurrency mining operations, adding another layer of high, often inflexible, demand on the ERCOT grid.
  • Other Emerging Hubs: Areas like Phoenix, Arizona; Atlanta, Georgia; Columbus, Ohio; and Hillsboro, Oregon are also experiencing significant data center growth, driven by factors like tax incentives, suitable land, water availability (for cooling), and network connectivity.
• Grid Challenges:
  
  • Pace Mismatch: Data centers can often be planned and built faster than the complex, regulated process of upgrading high-voltage transmission lines and substations needed to serve them. A single large data center campus can require hundreds of megawatts – equivalent to the power demand of a small city – creating step-change increases in local load rather than gradual growth.
  • Transmission Bottlenecks: Existing grid infrastructure in these hot spots wasn't designed for such concentrated, massive power draws. This leads to congestion on transmission lines, potentially limiting the ability to connect new facilities or requiring costly and time-consuming upgrades.
  • Reliability Concerns: The sheer scale of demand raises concerns about maintaining grid reliability, especially during peak hours or extreme weather events. Utilities worry about the potential for localized outages or the need for curtailments if supply and delivery capacity can't keep pace.
  • Utility Response: Energy providers and grid operators are exploring solutions ranging from accelerated transmission build-outs and grid modernization investments to implementing special tariffs for data centers that reflect their unique load profiles. Increasingly, there's discussion around demand response programs tailored for data centers or even requirements for on-site generation or energy storage to mitigate their grid impact. Early and close collaboration between data center developers and utilities is becoming essential to manage this growth sustainably.

Planning and opportunities for retailers 

How can energy retailers adapt and thrive in this environment? The surge in demand puts retailers in a tricky position; they are largely at the mercy of external forces, with limited ability to influence either demand or supply, except for larger vertically integrated energy companies. 

Nonetheless, the retailer that successfully navigates these choppy waters will  reap the rewards. 

One fairly obvious route to profiting from data center proliferation is through PPAs. There have been some major deals already, notably the restart of the Three Mile Island nuclear plant to supply Microsoft data centers. Some of these PPAs will be co-located, but there are still many, like Three Mile Island, that will involve retailers. Fortunately, Gorilla has an application for you.

However, PPAs are not the answer for every data center. Data centers are high-demand consumers with complex, fluctuating load profiles that don’t follow traditional patterns and won’t be able to find PPAs in every circumstance. For more traditional bespoke deals, retailers must be able to configure, price, and validate complex deals rapidly—sometimes in response to multiple RFQs in a single day. This demands pricing systems that are both highly automated and fully transparent, with the granularity to calculate at meter level and settlement interval.

Furthermore, traditional forecasting models—based on aggregated, static profiles—fall short in the face of data center volatility. Gorilla’s forecasting applications are designed for agility and accuracy, enabling energy retailers to incorporate granular usage data, integrate weather and renewable variability, and run high-frequency updates. This means forecasting models can stay in sync with the dynamic operations of data centers—reducing imbalances, avoiding penalties, and improving procurement accuracy.

Perhaps the biggest hurdle for many retailers is the fragmented nature of their systems. Pricing, forecasting, trading, billing, and settlements often rely on disconnected tools and data sources, slowing down decisions and introducing risk. A unified, interoperable data layer that connects the entire energy value chain enables alignment across teams, better insights, and faster innovation.

Beyond operational excellence, energy retailers will benefit from evolving their commercial strategies to reflect the unique characteristics of data center customers. This could include:

• Specialized rate plans that recognize the high load factors of data centers while building in incentives for reliability and grid-friendly behavior.
  
  
• Demand response programs or on-site generation partnerships, enabling data centers to contribute to grid stability while lowering their energy costs.
  
  
• Enhanced long-term forecasting models that incorporate large, stepwise demand increments when new data centers come online—ensuring upstream procurement, infrastructure planning, and portfolio strategies are ready to absorb the impact.

Meeting the needs of data centers is not just about scale—it’s about speed, precision, and adaptability. As the energy landscape becomes more complex, forward-thinking retailers are turning to modern, cloud-native platforms to transform their operations. Platforms like Gorilla, built specifically for the energy sector, offer a unified, real-time approach to forecasting, pricing, billing, and profitability—enabling retailers to serve the next generation of energy consumers with confidence.

‍