Now I'm building the intelligence layer above the grid.
SoftServe brought me in as NVIDIA Solutions Director for Power & Utilities to build GridCORTEX, the AI-powered utility intelligence framework for electric utilities. This is the role where everything I've built across the previous five chapters becomes the foundation for something the industry has never had: an intelligence layer that sits above the OT stack, answers the questions no system of record was ever built to answer, and now extends into the physical world with autonomous robotics and edge AI.
GridCORTEX is built on NVIDIA's accelerated computing platform, Earth-2 for weather physics, Isaac ROS 2 and Jetson Orin for physical AI, Omniverse for digital twins, NeMo and NIM for generative AI grounded in utility records, Metropolis for vision AI, and cuOpt for fleet-scale optimization. It encompasses 187 utility-specific use cases across 23 operating domains and ten independently deployable domain agents, entering through the four domains utilities are funding today: Connect, Resilience, Asset, and Operations.
The 33 years of sitting in utility control centers, building the platforms, integrating the systems, and advising at the partner level are what make this different from a technology company building utility AI. I have operated the grid that GridCORTEX is designed to augment. I know the questions that fall between the systems because I lived in the gap layer before it had a name.
What GridCORTEX is.
The intelligence layer above the grid: an AI framework that sits on top of the systems a utility already owns, answers the questions those systems cannot, and never touches the controls. Augment the grid, never replace it.
What I'm capturing here
AI solutions leadership, translating 33 years of operational and consulting perspective into the intelligence layer the utility industry has always needed but only now has the compute to build. The fifth clock, physical AI, is the newest and may be the fastest-moving. And I'm in the right seat to build for it.