Birds of a Feather Archive

AI Factory as the Unit of Compute: Driving Modularization, Standards, and Reference Designs

Authors: Greg Stover (Vertiv), John Gross (J.M. Gross Engineering, LLC; ASHRAE TC9.9), Martin Olsen (Vertiv), Ali Heydari (NVIDIA Corporation)

Abstract: This interactive panel will explore how AI factories—from 130 kW pods to gigawatt-scale campuses—are emerging as the unit of compute, requiring modular design, standardization, and reference architectures to enable rapid, scalable deployment. Panelists will share lessons learned implementing these diverse solutions, addressing power, cooling, and service strategies needed to democratize AI at scale. The session will encourage audience participation to discuss practical approaches for designing, deploying, and operating AI infrastructure of all sizes to meet growing demands efficiently and sustainably.

Long Description: This Birds of a Feather session will explore AI Factory as the Unit of Compute: Driving Modularization, Standards, and Reference Designs. As AI workloads grow exponentially, data centers are evolving into standardized AI factories—scaling from 130 kW pods to gigawatt-scale campuses. This shift demands modular design, reference architectures, and practical standards that simplify deployment, reduce costs, and accelerate time to value. Panelists will share lessons learned from designing, deploying, and servicing these facilities across diverse scales, highlighting both technical challenges and operational realities.

Attendees will benefit from the panel’s deep industry expertise - no team has more experience deploying AI infrastructure in last 24 months than this panel:

*Moderator Greg Stover (Vertiv) brings over 30 years in data center efficiency optimization, championing innovative power, cooling, and monitoring solutions tailored for AI and HPC needs, with experience supporting hyperscale, colocation, and edge environments.

*Wade Vinson (NVIDIA) is an industry pioneer with 20+ years of experienc in HPC design and deployment. Wade is currently NVIDIA's Chief Data Center Architect.

*Martin Olsen (Vertiv) has led global strategy and deployment for data center solutions with deep experience in modular design, thermal systems, and integrated planning across hyperscale and enterprise environments.

*John Gross (J.M. Gross Engineering) is a mechanical engineer with nearly two decades of experience in data center design and commissioning, specializing in liquid cooling, hyperscale and HPC deployments, and authoring industry standards with ASHRAE.

The highly interactive session will invite audience participation, with about 50% of the time to open discussion and audience Q&A. This session will deliver practical, vendor-neutral insights to help attendees plan, deploy, and manage AI-ready infrastructure of any size, promoting shared learning and industry alignment.

Key Talking Points: *Defining the AI factory as the “unit of compute” for AI workloads *Driving modularization and standardization to enable scalable, interoperable deployment *Lessons learned across design, deployment, commissioning and service phases *Best practices in power and liquid cooling integration for AI environments *Vendor-neutral approaches to democratizing AI deployment at any scale, at speed *Audience discussion of real-world challenges and opportunities for industry collaboration

Back to Birds of a Feather Archive Listing