The Department of Energy’s Pacific Northwest National Laboratory has partnered with Amazon Web Services to explore and validate some of the most advanced artificial intelligence tools that could help operators of the nation’s electric grid respond more effectively to severe weather, rising electricity demand and cyber or physical threats.
Next-generation AI systems
PNNL has extensive experience in developing advanced computing tools for grid operations, dating back to the 2003 blackout when they implemented machine learning to analyze data from grid sensors. A recent 2025 report highlighted the potential of AI, including large language models, for grid operations, while also noting the challenge of 'hallucinations' in LLMs. To address this, researchers are exploring neuro-symbolic AI, which combines pattern recognition with domain-specific knowledge like physics. AWS's neuro-symbolic AI tools will be tested to help operators make rapid decisions during disruptions such as wildfires or cyberattacks, aiming to augment human oversight without overloading operators with information.
Expanding, accelerating grid
The partnership with AWS is timely, as large-scale computing facilities, such as data centers, are causing a rapid increase in electricity demand and instantaneous changes in power usage. Existing grid protection systems and human operators are currently ill-equipped to manage these fast dynamics. The collaboration focuses on developing automated tools and AI-driven decision support to process complex information quickly and provide actionable insights for human operators. This initiative is part of DOE’s Genesis Mission, reflecting a broader national interest in building trustworthy AI tools to enhance grid resilience and secure critical infrastructure against a more dynamic and demanding future.
