A research team led by scientists from Austria argues that the rapid growth of AI data centers can accelerate the clean energy transition when paired with large-scale energy storage deployment. They introduce a novel concept – an “AI-energy storage nexus” – to describe a mutually reinforcing relationship between AI infrastructure and energy storage technologies.
“Unlike common expectations, our paper finds that AI growth is creating opportunities to put clean energy transitions on track,” researcher Behnam Zakeri told ESS News. “This is mainly due to the nexus effect between AI and energy storage. These two technologies are creating a virtuous cycle where the development and scaling of one technology has positive impacts on the development of the other. We identify where these synergies are happening, and for the first time we coin the term ‘AI-energy storage nexus’ for that matter.”
Zakeri said energy storage can resolve some of the big problems of digital infrastructure growth, not only in terms of renewable energy matching, but also in shortening access to the power grid and providing flexibility for the phased connection of data centers to the grid.
“We also show how the investments by big tech companies have shifted to energy storage solutions, which will advance research and development and deployment of such technologies, with spillovers for the broader energy transition beyond the digital infrastructure,” he said.
The researchers identified several ways the two technologies reinforce each other within the AI-energy storage nexus. They said machine learning and generative AI can accelerate the discovery of molten salts for thermal storage, identify new battery chemistries, and improve battery lifetime and recyclability. At the same time, rising demand from AI data centers is helping commercialize long-duration energy storage, with Google signing off-take agreements for liquid CO₂ and iron-air storage.
The researchers identified five main ways energy storage can support AI infrastructure. They said it can reduce peak grid imports, easing interconnection constraints and delaying costly transmission upgrades, while also improving power quality and reliability by providing backup power. Energy storage can also enable demand response, allowing data centers to reduce demand or export electricity during grid emergencies, and increase self-consumption of on-site renewable energy. In addition, it can enhance operational flexibility through workload scheduling, load balancing, and other forms of flexible computing.
The authors conclude by proposing four policy recommendations to ensure AI supports the clean energy transition. They call for reforms to grid connection rules that reward flexible, grid-supportive data centers, the removal of barriers to co-locating battery storage and renewable energy, incentives for grid-aware computing and demand-side flexibility, and targeted procurement and financing policies to accelerate long-duration energy storage deployment.
“We plan now to develop models for the operation of energy storage co-located with data centers or in the near grid that would resolve some of the challenges facing the power grid today,” added Zakeri.
Their article, “The AI–energy storage nexus: opportunities for clean energy transitions,” was recently published in Energy and Climate Change. Researchers from Austria’s Vienna University of Economics and Business (WU), the International Institute for Applied Systems Analysis (IIASA), and Saudi Arabia’s King Abdullah University of Science and Technology contributed to the paper.
The post Researchers outline AI-energy storage nexus for clean energy transition appeared first on Energy Storage.
