Vertiv and Deloitte Center for Integrated Research have introduced a framework to help enterprises determine where artificial intelligence (AI) workloads should run and whether their existing physical infrastructure can support production deployment.
The Vertiv Designing Infrastructure for Enterprise AI Guide, developed by Vertiv and digital infrastructure research firm STL Partners, provides business, IT, and facilities leaders with a workload-based approach to planning AI infrastructure. It helps organizations assess power, cooling, connectivity, and operational capacity before committing to deployment.
As businesses move AI projects from experimentation to production, infrastructure readiness is becoming a critical factor in determining deployment locations and implementation timelines. Depending on their requirements, workloads may run in public cloud, colocation facilities, enterprise-owned infrastructure, hybrid environments, or edge locations.
The guide outlines a four-step process: define the business outcome and AI workload, assess its scale and performance requirements, determine the appropriate deployment location, and match the workload to a suitable physical infrastructure model.
It also identifies six enterprise AI workload archetypes covering model development, retraining, and four categories of production use. These are matched with four infrastructure models: Edge AI Cluster, Small AI Cluster, Large AI Cluster, and High-Density AI Cluster.
Each model provides a planning reference for power, cooling, connectivity, resilience, operational support, and future expansion. Organizations can adapt the corresponding reference designs following site assessments and detailed engineering validation.
Vertiv CEO Giordano Albertazzi said enterprises must consider the power, cooling, connectivity, and operating environments required by individual AI workloads to establish a viable path to production.
Deloitte Consulting LLP principal Nicholas Merizzi emphasized the importance of evaluating business value, governance, operational ownership, and infrastructure together as AI workloads evolve.
The guide also includes a four-phase roadmap for planning, deploying, and evolving enterprise AI infrastructure, helping organizations identify facility limitations early and make more informed technology investment decisions.
The full guide is available at Vertiv’s official website.
