Global renewable energy capacity increased by 50% in 2023. At this pace, the COP28 target of tripling capacity by 2030 potentially seems attainable. However, sustainable energy systems are about much more than just capacity - energy must reach the right people, at the right time, and the variability of renewable sources and peak demands make this a critical challenge to confront. Thus, there remains a critical limiting factor not discussed nearly often enough: reliable, affordable, distributed, and resilient energy storage. Efforts to scale and innovate in energy storage must intensify especially to match the accelerating demands of fast-growing industries such as electric mobility and utilities. 

In our previous article, we explored how digital technologies are transforming the entire energy value chain. Now, let's delve into the specifics of one crucial aspect: energy storage. 

AI-managed storage can tip the economic equation to unlock investment in renewable energy

The technical potential of renewables in developing countries is immense – 3.1 TW just from offshore wind alone. The bounds are created primarily by how quickly we can scale the infrastructure to convert and deliver this energy to people and industry. Although there are essential links to climate adaptation and fundamental human development, ultimately the ability to scale these systems becomes a question of economics, for investors to be able to justify the costs of investing in renewable systems. 

Artificial Intelligence (AI) offers significant potential to offer integrated advancements and optimized systems across the energy storage value chain, which can shift investment potential in renewable systems in places it is needed most. 

First, we must optimize our existing energy storage and delivery systems with AI

Before we discuss the many potentials for AI in deploying new systems or discovering new technologies, its critical utility lies in improving our existing infrastructure through data optimization.

Many developing countries have unreliable grids with frequent power outages. Improving grid conditions in these regions is important, as the investment towards them is decreasing over the years resulting in a lower quality of grid connection, and only adding more input from renewable energy sources would not solve the issue. Here, AI-driven energy storage could be a potential solution to solve this grid connection challenge  by enabling better integration of renewable energy sources with the grid and ensuring grid stability. 

AI algorithms can handle vast datasets in real-time from various sources, extensively analyzing energy demand, grid conditions and environmental factors to dynamically adjust the charging and discharging of storage systems. This ensures that energy storage assets are utilized at their optimal capacity, reducing wastage and maximizing overall system efficiency. This approach enables more sophisticated management of grid-scale energy storage, helps prevent fluctuations in energy supply and demand and enhances grid stability. Evergen is an example of an AI-driven platform designed to maximize the utilization of solar and battery energy resources. It achieves this by avoiding peak demand, offering demand response services, and managing grid stabilization.

Second, we need to accelerate decentralized energy access through AI-enabled microgrids

AI is making possible new horizons for developing and managing microgrids in remote areas. With AI, these microgrids can enhance distributed renewable energy by autonomously managing local energy production, storage, and distribution, tailored to local conditions without constant human intervention. These self-contained grids with local generation and storage can provide electricity to off-grid communities, improving access to power for lighting, communication, and essential services. This democratized model of energy access also allows small-scale renewable producers and consumers to participate in energy markets. 

Recognizing these benefits, new initiatives are emerging to leverage this potential. In Sri Lanka, the Asian Development Bank is developing a framework for the application of AI, facilitating the integration of more distributed renewable Energy into the grid without compromising stability. At Florida International University (FIU), researchers are using AI to help microgrids predict grid instability, allowing them to isolate before outages. This proactive approach ensures continuous power for buildings on the campus, effectively creating a smart city environment.

Third, we should leverage big data and AI for integrated benefits across sectors

Recent research has shown that almost 60% of healthcare facilities in 46 low- and middle-income countries experience unreliable power. By deploying AI-integrated energy storage systems, these critical facilities can benefit from a reliable power supply for essential medical equipment, such as refrigerators for vaccines and lighting for life-saving operations, significantly improving healthcare delivery in remote areas. 

Furthermore, AI-enabled storage systems enhance resilience by predicting grid stability, which is especially beneficial during extreme weather events like high heat, cyclones, and thunderstorms that can destabilize the grid. In this context, Small Island Developing States (SIDS) present ideal testbeds for piloting the application of AI-enabled energy storage systems, given their unique circumstances. For example, the Faroe Islands have pioneered the use of a virtual power plant and smart grid.  

Fourth, we must create an environment for innovation

Developing innovative energy policies that incorporate AI technologies requires interdisciplinary and multi-dimensional decision-making, considering factors such as energy type, scale of implementation, AI methods, and automation levels. This approach must be complemented by bottom-up efforts to build technical capabilities and develop digital tools that facilitate transformative use and innovation on the ground. 

As AI models heavily rely on data inputs for decision-making, any biases in the data then can lead to inadequate outcomes, making the “garbage in, garbage out” concept crucial for input data quality. Creating policies that allow open data and collaborations between researchers, government agencies, and businesses would help to accelerate data sharing and develop more effective AI models. Besides that, governments can offer incentives such as grants and subsidies to drive collaborations across sectors such as AI companies, energy firms, and research institutions to drive AI-driven innovations for energy storage.  

Furthermore, concerns about data security and privacy emphasize the urgent need for robust encryption protocols and access controls to ensure the privacy of sensitive energy information. Additionally, ensuring interoperability across different technologies is also critical to enable seamless integration within the energy systems. The ongoing research on federated learning and edge computing is one of the potential promising solutions to enhance data security while allowing collaboration across energy networks. 

A government-led approach by creating a supportive policy and regulatory frameworks for AI to be further implemented in the energy storage industry is crucial. Alongside, it is critical to solve concerns pertaining to unauthorized access, manipulation, or even disruption of energy storage and distribution systems.

While leveraging AI is crucial, it is equally important to address broader systemic issues such as existing socio-economic disparities, policy barriers impeding equitable energy access, and infrastructure inadequacies which limit to effectiveness and scale of AI solutions in energy storage.