How AI Is Powering the Next-Gen Electric Grid

Key Takeaways

• AI Becomes the Core of the Smart Grid
  Artificial intelligence is now integral to power-grid operations, advancing the original smart-grid concept into the emerging AI electric utility.

• Massive Market Growth
  The global smart-grid market — especially AI-driven applications — is projected to reach $208 billion by 2032, reflecting strong worldwide adoption.

• Real-World Deployments

  • CAISO & OATI Pilot: First ISO to test a generative/agentic AI platform for outage management and control-center operations.

  • Duke Energy: Integrating AI across the enterprise, from predictive maintenance of transformers to reducing interconnection study times.

  • EPRI Open Power AI Consortium: Over 116 industry leaders collaborating to accelerate AI solutions for reliability, efficiency, and customer experience.

• Beyond Automation
  AI is evolving from a tool for efficiency to an intelligent collaborator, blending machine insights with human critical thinking to detect issues and optimize grid performance.

• End-to-End Digitalization
  AI links distributed energy resources, asset management systems, demand-response programs, and vegetation-risk platforms, enabling a fully self-aware grid.

• Risk and Resilience
  Utilities are risk-averse, but proven AI applications are now viewed as essential for reliability, resilience, and climate-change adaptation.

Who would have thought when smart grid technology was introduced it would have had the impact it has? Initially, it was considered more of a hodgepodge of digital technologies than anything approaching a smart grid. Integrating sensors, transducers, and controllers into critical equipment like large power transformers to monitor the essential parameters made sense to many stakeholders. Large transformers are expensive, have long lead times for replacements, and impact a utility’s ability to deliver electricity. Something like a smart electric meter, however, was a different story and acceptance took a while.

The technology improved with features like two-way communications and control, which added to the appeal of the smart grid concept. The availability of real-time data proved valuable for monitoring a wide range of utility assets and bringing self-awareness to both sides of the meter. Smart grid applications got a needed boost with the introduction of UIoT (utility internet of things), a branch of IoT (internet of things) technologies. UIoT dealt specifically with technologies within the utility sector, and it brought attention to their benefits. Several decades later smart grid technology has become so intertwined with the power delivery system that it’s impossible to have one without the other.

The worldwide marketplace has responded favorably to this condition, but it has been the influences of artificial intelligence (AI) that’s getting everyone’s attention. According to an “AI Overview” query, smart grid’s global market has undergone a positive uplift with the integration of AI (artificial intelligence) into its digital-based components. The inquiry indicated that smart grid technologies, particularly those driven by AI, are experiencing robust growth. It’s expected that this market will reach an estimated US$208.15 billion by 2032.

So What's Next?

AI-powered digital systems are transforming the power grid’s landscape by providing it with the self-awareness needed for the dynamic environment the power grid finds itself in today. A few months ago we looked at the holistic approach of the end-to-end digitalization the power grid needs to keep up with our rapidly changing landscape. This growing trend is needed to improve the power grid’s performance, which brings us to the next logical step. The AI electric utility is being seen as the most effective approach to consolidate AI and the smart grid.

What is an AI electric utility? Defining that isn’t easy with an undertaking that’s still developing and evolving. Like so many of today’s digital-based applications, the definition is somewhat fluid. Let’s just say it’s the optimal method to combining AI-driven functionality with the smart grid applications. As one group put it, the goal is to upgrade the operations of the existing electric power grid. A DOE (Department of Energy) report points out that advancements in AI are maximizing energy operations while reducing risks.

That’s pretty much the goal of every sophisticated digital platform that’s been deployed by the utilities and grid operators in the past several years, but the AI utility is a bit different. We have a wide selection of advanced technological applications that watch, supervise, and manage everything from vegetation dangers to power flows, and the number is growing. Merging them provides a critical awareness that’s best achieved when these systems are working together rather than continuing as individual elements.

Operating Beyond Old Boundaries

An illustration of this is the asset management system’s (AMS) evolution. AMS has been moving in this direction since its inception. As AMS matured, it produced really big-data that needed AI to turn it into useful information. Basically, that led to asset health modules, which brought about performance modules. Other useful elements came along that had positive influences on other digital systems, which in turn grew in numbers. As the old adage says, “A rising tide lifts all boats.”

Schemes such as distributed energy resource management systems, demand response programs, vegetation risk platforms and others followed this same process. Features were added and functions led to interactions between the various smart configurations and AI became the glue holding it all together. AI-driven functionality linked multiple databases and integrated them with a variety of digital asset technologies. Interestingly, we are just starting to see how the AI electric utility fits into this process and some other more well-known applications.

“Charging Ahead” has featured many of these AI-enhanced tools this year such as the virtual power plant, community microgrids, battery management systems, etc. These unconventional approaches symbolize more than a band-aid for the existing power grid. Each is a piece of the AI Utility’s virtual infrastructure needed by the aging infrastructure to address the growing challenges of climate change and the grid’s growing power demands. Let’s look at actual advanced AI utility projects

Deploying the Next Generation

A few months ago, the California Independent System Operator (CAISO) and Open Access Technology International (OATI) announced a groundbreaking pilot program that the two have entered. It marks the first time that an ISO is deploying AI technology to streamline operational processes starting with outage management. A press release from OATI said, “It was the world’s first generative and agentic AI platform purpose-built for the energy industry.” CAISO is the balancing authority for 80% of the state of California and parts of Nevada. CASIO is also the operator of the Western Energy Imbalance Market, so this is a major generative AI pilot project.

The release continued saying, “CAISO’s pilot will determine the extent to which its outage management system, which includes structured tables, unstructured text and long review times for system operators, can be made nimbler and more efficient using generative AI.” CAISO stated this program fits perfectly with their ongoing control center modernization program. It was expected that this tool would improve the operators’ situational awareness and free up time for other important tasks that can make a real difference for those operators. Testing is expected to start later this year, which will move the AI utility another step forward.

Earlier this year, the AI utility received backing from EPRI (Electric Power Research Institute) in the form of its “Open Power AI Consortium.” EPRI’s announcement stated the Consortium membership numbers have over 116 industry leaders (at the time of this writing). They consist of electric utilities, technology providers, and researchers from around the world.

EPRI said that the Consortium, “will drive the development and deployment of cutting-edge AI solutions tailored to enhance operational efficiencies, increase resiliency and reliability, deploy emerging and sustainable technologies, and reduce costs while improving the customer experience.” EPRI continued saying the Consortium “aims to evolve the electric sector by leveraging advanced AI technologies to innovate the way electricity is made, moved, and used by customers.”

Also, Duke Energy has released several announcements this year about their plans for deploying AI throughout their operations. It’s said to be a holistic integration of AI-powered technology across the entire organization ranging from business development to maintenance and everything in between. One report said Duke would be using generative AI to reduce interconnections study times from weeks to minutes. Another release said Duke was combining human expertise with AI’s predictive maintenance to improve operations and the health of critical equipment like power transformers. This is tapping into the AI electric utility big-time!

AI Collaborator 

These are only a small portion of news releases and announcements coming from utilities and grid operators telling how they are using AI-driven technologies. Overall it appears that AI in general and generative AI in particular are finding their way into the power grid in growing numbers. Accenture’s “Tech Vision 2025” said, “AI is no longer just a tool for efficiency; it is becoming an intelligent collaborator, capable of reshaping entire industries.” Surveys are listing percentages of utilities utilizing AI ranging from 30% to over 70%, but it’s the growing interest that’s important.

Overall, some utilities are totally embracing AI, while others are in the initial stages of acceptance. Most users, however, reside in between, which isn’t surprising. The power delivery industry is more risk-averse than other industries because keeping the power flowing is critical to modern society. Still it’s important to keep in mind that many AI technologies have moved into the tried-and-tested category of the smart grid’s tool set, which helps avoid risk. Utilities who are using these sophisticated digital technologies have tremendous advantage over those who are not using them.

Interestingly, there’s a subtle shift that kept popping up in the research for this article. It’s the blending of AI’s abilities with human expertise that was discussed by many utilities in their narrative of deploying AI on their systems. Developers found that humans brought innovative problem solving into the relationship. In addition, humans can identify potential blind spots that AI is unaware of. They recognize a possible problem through critical thinking that the AI applications lack. This shift is beginning to look like a trend, which will be interesting to watch as it progresses. Ultimately AI utility needs people and AI technology working together to be successful!