- Home /
- Summit Series 2026 /
- Tektronix /
- Sessions /
- From Grid to GPU: Validating Next-Gen AI Data Center Power Infrastructure
From Grid to GPU: Validating Next-Gen AI Data Center Power Infrastructure
Shifting to 800V HVDC architectures introduces complex high-current challenges. Discover automated measurement workflows and simulation techniques to test and debug rack-level distribution.
AI workloads are pushing data centers beyond traditional AC rack distribution toward 800V HVDC architectures, reshaping how engineers design, validate, and debug rack-level power systems.
In this presentation, we’ll explore how next-generation power systems are tested, debugged, and simulated using oscilloscopes, regenerative power supplies, high-power electronic loads, and automated test systems. We’ll cover practical measurement challenges across power shelves, busbars, BBUs/CBUs, VRMs, and rack-level PDNs, including transient response, ripple and noise, dynamic loading, impedance, efficiency, and timing-correlated power integrity analysis.
You’ll also see real-world examples of using regenerative systems to simulate AI rack behavior and validate modern data center power infrastructure.
Whether you’re designing converters, validating AI power systems, or debugging high-current PDNs, this session focuses on the measurement techniques and automated test workflows engineers are using to develop the next generation of data center power.
Share
Time Left
Happening now
About
James “Bald Engineer” Lewis is a Tektronix Senior Technical Marketing Manager focused on power conversion and power delivery networks (PDNs) in modern data center architectures. His work covers high-voltage front ends, including 800V distribution, emerging 48V rack-level systems, and board- and chip-level PDNs, where efficiency, stability, and transient performance are tightly coupled.
James has helped engineering teams debug noisy rails, fix probing errors, and see how their PDN behaves in real conditions. Whether validating high-power conversion stages, characterizing rack-level behavior with dynamic load and source emulation, or correlating system performance to on-board PDNs, he translates complex power integrity challenges into practical measurement workflows that engineers can apply immediately.
