This study investigates whether structural fragmentation in electricity networks can provide early warning signals for extreme price spikes. Using one year of PJM real-time electricity market data, the analysis examines cross-sectional dispersion of locational marginal prices (LMP) across network nodes. The results show that higher network dispersion is associated with substantially higher probability of next-day extreme price spikes, even within the same system price regimes. These findings suggest that structural signals inside infrastructure networks may provide early indications of systemic stress before instability becomes visible in aggregate market prices.