Venkata Kondepati, PMP, is Manager of Data Architecture & Engineering at Ascentt, specializing in Cloud, Data, AI, Customer IAM and GIS.
getty
For many decades, electricity and power traveled in a linear direction: produced at a power plant and distributed to homes and businesses. The demand was predictable, and power companies could adapt relatively easily. However, various factors that increase power demand or disrupt stability are straining the existing model. The main factors affecting electricity consumption are the advent and growth of electric vehicles, extreme weather events, the increasing number of data centers and the distribution of energy resources.
As outlined in the International Energy Agency’s Electricity 2026 report, it is clear that the demand for global electricity is rapidly growing. Previously, the challenge in most scenarios was simply producing enough power to meet increasing demand, but now it is more important to deliver that power both efficiently and intelligently across a traditionally designed grid. Specifically, the existing grids aren’t designed to handle significant variability. An effective solution to this challenge is GeoAI, which combines geospatial intelligence with large language models (LLMs) and machine learning to make a meaningful difference.
To truly dissect the problem with modern grids, it is important to understand their localized nature. While a system may appear effective and relatively stable, its individual components may experience stress from factors such as clustered electric vehicle charging, rapid housing development or even massive commercial loads. Recent research from the National Renewable Energy Laboratory (NREL) suggests that the increasing adoption of electric vehicles can worsen thermal and voltage constraints on distribution systems, further highlighting the need for a management system capable of handling the load. Thus, the main question is “Can a specific transformer or feeder support future demand?” given the varying nature of the grids.
Planning tools that were traditionally used cannot provide an effective, concrete answer because the data they use is gathered from an array of sources, including GIS, SCADA, AMI and asset systems, which are challenging to integrate into a clear vision. GeoAI, however, can integrate data from all these sources to provide a unified operational view.
The main element that defines how effectively a system functions is structure. The ArcGIS Utility Network provides a geospatial model that highlights the connections between assets, whether between substations and feeders or transformers and customers. According to ESRI, the utility network is crucial for enabling connectivity modeling and tracing, which allows utilities to understand how electricity flows across the grid.
The main distinction is that instead of:
“Demand will increase by 20%.”
Utilities will be able to give clear information that is actionable, such as:
“Feeder 12 will exceed capacity at 7 p.m., under current projected EV growth.”
The difference between these two results is astounding because it means the difference between effectively meeting demand and failing to adapt the grid systems.
It is unrealistic to build entirely new infrastructure for both cost and efficiency reasons. Instead, optimizing existing infrastructure can help mitigate transition costs. GeoAI is powerful because it allows utilities to perform the following:
• Shift EV charging power away from peak hours
• Execute batteries during constrained periods
• Focus demand response programs geographically
• Delay expensive upgrades
The key is not to reduce overall consumption, but to optimize so that demand can be handled across different times and locations. Additionally, advanced analytics also need to be usable and easy for humans to interpret. LLMs are crucial for translating complex grid data into plain language for planners and operators. For instance, a result such as:
“Transformer T-204 is at 92% capacity due to EV charging. Shifting 20% of the load to off-peak hours eliminates overload risk.”
The next step is using agentic AI, which will monitor grid conditions, coordinate across tools and trigger actions. These systems can detect rising feeder stress, identify flexible demand and even recommend or initiate load-shifting actions. The U.S. Department of Energy has noted that AI has the potential to improve grid reliability and detect anomalies, underscoring the importance of safeguards for critical infrastructure.
While GeoAI presents many advantages and technical possibilities, there are also novel risks that will need to be accounted for. For instance, data quality is important because gaps in the foundational data can lead to poor decisions. Additionally, cybersecurity is pivotal because the complexity and size of interconnected systems warrants heightened security given their exposure. Ultimately, the benefits need to be distributed fairly across communities to uphold equity.
One strategy to combat these risks and potential problems is human-in-the-loop decision-making, so that solutions can be cross-checked before implementation. Additionally, strong data governance and robust cybersecurity controls will maximize efficiency without problems. Audit trails for recommendations will further increase the benefits of using GeoAI.
Overall, the electric grid is on the path to becoming a dynamic system that will include capabilities such as real-time digital twins, dynamic hosting capacity for DERs and coordinated management of EVs, batteries and solar. Through these systems, energy providers can transform from providers to orchestrators of the ecosystem, maximizing output.
The challenge in the future will not be producing more electricity, but rather delivering it reliably and efficiently to areas of high demand. GeoAI, powered by ArcGIS Utility Network, is a feasible solution that optimizes the existing grid infrastructure for an energy-conscious future.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。