Connect | Virginia // Panel: Speed-to-power for AI factories - Rewriting the infrastructure playbook for GW-scale delivery

Secure your pass
  • —
  • Spotlight Stage | Lansdowne Ballroom
  • Energy & Sustainability
  • Mission Critical Power
  • Open Session
Fluence logo

Speakers

Sai Somayajula

Principal Electrical Design Engineer, NVIDIA

John DeBoer

Head of Data Center Vertical, North America, Siemens

TJ Winter

VP, Emerging Segments & Intelligence, Fluence Energy

As next-generation AI clusters scale past 100 MW toward multi-gigawatt campuses, power availability—not compute availability—has emerged as the critical bottleneck. Traditional sequential buildouts are no longer viable; the extreme power densities and millisecond-level dynamic load ramps characteristic of modern AI workloads require that power infrastructure, energy storage, and silicon architectures be co-engineered from day one.
In this panel, infrastructure leaders from Siemens, NVIDIA, and Fluence, convene to explore the technical and commercial framework required to accelerate energization and support reliable AI scaling. The discussion deconstructs the transition from standard data center design to integrated "AI Factories," examining how modular switchgear, battery energy storage systems (BESS), advanced thermal cooling interfaces, and silicon-informed facility topologies work in concert to unlock speed-to-power, mitigate severe transient dynamics, and enable energization sooner.
The panel will explore:
  • Ways to accelerate deployment timelines and overcome multi-year utility interconnection queues
  • Strategies to manage supply chain pressures and long lead times for critical power equipment
  • Understanding the collaborative blueprint required between silicon designers, power OEMs, and energy storage providers to scale GW-scale AI factories