Research - Utility AI Insights
April, 2024
A new technological age is underway in the power sector. Utilities, tasked with playing a leading role in how the economy decarbonizes, are increasingly looking for digital solutions to complicated problems. The magnitude of complexity is staggering — manage an energy and technology transition while maintaining the imperative of resiliency, reliability, and affordability. But improved sector digital maturity, buoyed by the availability of multi-modal data, is helping to meet an uptick in weather events, distributed energy resource (DER) deployments, and evolving customer needs. Increasingly, sophisticated artificial intelligence solutions and software are advancing the energy transition. In new research, and in partnership with Latitude Media, we delve into this space and provide a temperature check, opportunity guide, and a capability forecast for a rapidly changing sector.
Converging Power and Technology Trends
For more than 40 years, two converging macro trends have shaped the utility sector and initiated an inflection point. These technology and industry specific trends are accelerating and enabling transformation across the power sector. As we enter this new technological age in the sector, AI may provide some of the solutions the sector needs. The growing availability of data from years of sensor and digital infrastructure deployments and the increasing capabilities of utilities in the data science domain are giving power companies the chance to solve pressing problems while creating opportunity. As a result, utilities are now looking to leverage new and existing AI capabilities.
AI isn’t new to utilities. Many have been working with load forecasting and other machine learning applications for some time. But the extent to which AI is being incorporated into grid systems has increased dramatically in recent years with the proliferation of cloud-based solutions, the drop in chips prices, and the evolution of utility digital programs. AI offers up a true inflection point across the sector. For prepared utilities, AI has the power to accelerate ongoing digitization efforts.
““We are getting more accurate, more quantitative, and more insightful. For example, being able to predict the reliability improvement of a particular activity or evaluate the different options that serve load. If we see EV penetration in an area in the future [we can model] the impacts on a particular circuit, substation, or transformer - basically whatever the insight we need.””
AI Capabilties are Increasing Application Value
Various applications are benefiting from an increased sophistication in AI capabilties. Solutions such as predictive analytics for demand forecasting and preventative maintenance, image recognition for drone or satellite-based inspections of infrastructure and automation for routine tasks such as meter reading or grid management have all benefited from a rise in capabilties.
These capabilities benefit from the availability of AI frameworks and packages that encompass a range of software libraries, tools, and frameworks that make the design and implementation of AI solutions more manageable. Libraries such as TensorFlow, PyTorch, NLTK, and OpenCV facilitate data preparation, training/testing AI models, and help with deploying these models for practical use. Additionally, sophistication in ML, deep learning, natural language processing, and computer vision help to develop possibilities across the evolving application layer. The hardware layer is improving too, with high-performance servers, GPUs, and large-scale storage systems handling vast data sets from smart meters, sensors and various internal and external sources. These powerful processing capabilities are enabling real-time analysis, predictive modeling, and decision-making.
Utilities Launching New, and Enhancing Existing Use Cases
Non-critical and ‘narrow AI’ utility applications are being piloted across the value chain. Emerging patterns within the power sector suggest that AI is a double-edged sword for utilities, offering both significant opportunities and risks. There is an appetite for multi-model and multi-use case AI solutions that can integrate various forms of data—text, images, and sound—into a single AI model. Similarly, broad use cases that touch multiple applications appeal to utilities and investors. For example, vast libraries of related customer and grid data for leveraged insights.
Utilities increasingly see AI as an extension of the progressive trends in smart grids, data analytics, and digital transformation. Utilities are universally identifying and investigating diverse AI applications tailored to their specific requirements while mitigating associated risks. Many are engaging in sandbox testing to ensure smooth, wider-scale deployment. There is a strong emphasis on adopting a human-centric approach. This involves dedicated training including AI general sessions, vendor-led training, and upskilling programs for employees to adeptly manage AI technologies. Utilities often synchronize these initiatives with strategies to restructure roles and enhance leadership synergy.
As AI is driving down the cost of innovation thanks to code-generation tools like Copilot and the availability of robust open-source LLMs we expect to see solutions be implemented at an unprecedented pace over the coming years.
Utility Archetypes of AI Adoption
While a significant number of utilities are actively experimenting with AI, digital maturity and, in some cases, structural and regional attributes of a company lead to variances. The evolution within these archetypes suggests a sector that is responsive to both internal organizational factors and external technological advancements. As such, strategic AI adoption requires a nuanced approach that considers the unique position and context of each utility.
In the research, our "Utility Archetypes of AI Adoption Model" delineates the varied pathways and progressions of AI integration across the sector, providing a nuanced framework for understanding the strategic and operational maturity of different utility companies and their AI development phases.
Forward Looking Insights
Overall, the report assesses how the power sector is presently evaluating and adopting AI and explores the solutions that may impact its future. To paint this picture, we examine both the supply side and the demand side of the Utility AI market. Through interviews with multiple utility leaders, the assembly of numerous data sets, and by leveraging on the ground deployment experience, the research offers unique insight into the space.
Learn more about the research, over on Latitude Media, listen to this podcast below for further insights, and contact us to learn more about our power sector AI assessment, strategy and deployment capabilties.