Direct Air Capture has the potential to help meet the world's net-zero target Green hydrogen Developing membrane-less technology for hydrogen production is critical for the technology’s ability to scale. Researchers from Harvard, EPFL, and HPE have developed AI-enabled digital twins for membrane-less electrolyzers, with the potential to boost efficiency by up to 20% and cutting costs by up to 25%.67 Direct Air Capture (DAC) Researchers from Microsoft, MIT, and UC Berkeley are developing an AI model to accelerate the discovery of new materials for cost- effective carbon dioxide removal.68 The AI-driven model aims to identify materials capable of capturing large amounts of CO2 from the atmosphere while minimizing the energy required to release it for storage or reuse. Next-gen battery materials Battery storage faces challenges due to rising costs and limited supply of critical minerals like lithium. AI is helping to address these constraints. Microsoft and Pacific Northwest National Laboratories used AI to discover new battery materials with reduced lithium dependence in weeks instead of years.69 Small modular nuclear reactors (SMRs) Researchers at Purdue University and Argonne National Laboratory have developed an AI algorithm that predicts changes in SMR performance with 99% accuracy, potentially lowering costs and improving reactor efficiency, which would make nuclear energy more viable and easier to manage.70 These efficiency improvements and cost reductions are anticipated to make these technologies more economically viable, potentially attracting more capital and hence more adoption. Nuclear fusion At Princeton Plasma Physics Laboratory, AI is accelerating nuclear fusion development by predicting and mitigating plasma instabilities, optimizing complex computational tasks, and reducing computation time from tens of seconds to milliseconds.71 These advancements are key to making fusion a scalable and reliable energy source. Storage Octopus Energy’s Kraken platform is a powerful deep- tech solution designed to optimize the management of energy resources, particularly for energy storage systems. It leverages advanced analytics, real-time data monitoring, and machine learning to balance supply and demand, especially with the integration of renewable energy sources. AI’s game-changing ability to accelerate innovation shows particular promise in six key areas: AI has the potential to support a just and equitable transition to a net-zero energy system. To maximize its impact, AI must develop its capabilities and overcome obstacles to implementation. For example, the integration of AI into permitting processes will require improvements in the reliability and transparency of GenAI, significant process changes and the upskilling of personnel. Moreover, AI is not a silver bullet. Alone, it cannot address all the challenges associated with the Conclusion energy transformation. For example, any new materials for direct air capture will still need to be commercialized and funded at scale. Nonetheless, given the scale and urgency of the climate challenge, it would be remiss not to comprehensively explore and fully exploit the potential of AI. 42 Powering Possible 41