Prithpal Khajuria

De-risking Grid Integration for GW-Scale AI Infrastructure | Co-founder, vPAC Alliance (35+ Utilities & OEMs) | Intel

San Francisco Bay Area

About

Frontier AI scaling fails quietly — not in model collapse, but in infrastructure fragility. I work on the grid engineering layer that determines how fast AI infrastructure actually energizes. The protection coordination bottleneck between procuring power and delivering it safely is the constraint I've spent a decade solving across 30+ countries and four major U.S. ISOs (MISO, PJM, CAISO, ERCOT). In 2019, I funded and led the world's first virtualized protection relay demonstration in partnership with Southern California Edison and Salt River Project — validated within the IEC 61850 framework. That work became the foundation for co-founding the vPAC Alliance in 2023 — now a coalition of 35+ utilities, OEMs, and technology companies including ABB, Siemens Energy, and GE Vernova. At DistribuTECH 2026, three major OEMs announced commercial products based on the architecture. This is no longer a concept. It's an emerging commercial reality. The core thesis: every AI data center that self-powers, operates behind the meter, or co-locates with generation faces a protection coordination challenge that traditional utility frameworks aren't structured to handle at scale. Software-defined protection compresses the sequential engineering bottleneck that currently adds 12–18 months to interconnection timelines for loads above 200 MW. As AI workloads push campus power into gigawatt territory with increasingly dynamic load profiles, the intersection of grid-side protection and compute-side power behavior is where the hardest unsolved problems live. At Intel, I lead the Energy & Utility segment globally, managing strategy across utilities, OEMs, and infrastructure partners in 30+ countries — where I've built two industry coalitions from scratch and created a market category that generated $300M+ in ecosystem revenue.

Experience