NVIDIA outlined how energy and infrastructure companies are using its DSX platform to model, coordinate and manage AI-factory power systems.
NVIDIA described partner projects built around its DSX AI Factory Platform in a post dated October 21, 2026, positioning the software as a way to coordinate power-grid planning with AI data-center design and operations. The company cited a U.S. backlog of more than $300 billion in equipment and over 200 gigawatts of projects waiting in interconnection queues as evidence of infrastructure constraints.
The DSX suite combines tools for modeling facilities, adapting computing loads to grid conditions and coordinating data-center systems. NVIDIA says DSX Sim and SimReady assets let engineers simulate grid, substation and AI-factory components—including generators, switchgear, power systems and cooling—before construction. DSX Flex is intended to align facility demand with grid conditions and on-site generation, while DSX MaxLPS and DSX Exchange address compute efficiency and the sharing of operational signals across systems.
The post highlighted work with energy and infrastructure companies. GE Vernova is contributing grid-to-facility and rack-power expertise. Hitachi Energy’s battery energy storage system has qualified as NVIDIA DSX Ready, a designation tied to the company’s reference-architecture requirements. Schneider Electric, AVEVA and ETAP are working on power and cooling reference architectures for NVIDIA Vera Rubin and an Omniverse-based digital twin. Siemens Energy is developing a grid-aware power-infrastructure reference design aligned with the Vera Rubin AI-factory design.
NVIDIA also cited two examples focused on operating capacity. Phaidra integrated DSX Flex, DSX Max-Q and DSX Exchange into its self-learning agents for cooling and power allocation; NVIDIA says the setup delivered about 10% more compute within the same power envelope. Emerald AI integrated its Conductor platform with DSX Flex in an approach modeled to unlock up to 100 gigawatts on the existing U.S. grid. Those figures are claims and projections reported by NVIDIA, not independent measurements presented in the post.
The announcement frames power coordination as an engineering challenge spanning utilities, facility equipment and computing systems. The named collaborations cover simulation, digital twins, storage and workload flexibility, but NVIDIA’s post does not provide deployment timelines or independent validation of the projected grid capacity.