Vertiv announced a digital twin capability for AI factory infrastructure that lets operators simulate power, cooling, and deployment before building, addressing the gap between accelerator chip advances and the physical systems needed to support them.

What happened: Vertiv announced on June 1 a production-grade digital twin for SmartRun integrated into NVIDIA Omniverse DSX Blueprint, designed to help design, simulate, and validate power, cooling, controls, and deployment workflows as a single system before build-out. The company also introduced Vertiv PurgeRite NearZero on June 3, a fluid management service for data-center hydronic systems that reduced water consumption by up to 78% and discharge management costs by as much as 34% in selected deployments.

Why it matters: High-speed interconnects and dense accelerator clusters only work if the surrounding power and thermal infrastructure can support them at scale. Vertiv's digital twin is positioned as the first phase of a multi-phase roadmap to narrow the gap between accelerated compute innovation and physical infrastructure readiness, making AI factory infrastructure more configurable, repeatable, and simulation-ready.

What to watch: The SmartRun digital twin is the first phase of a multi-phase AI factory digital twin roadmap, signaling that data-center infrastructure—not just chips—is becoming central to how companies deploy AI at scale.