PJM Data Center Study → Interactive → Vulnerability Surface

Explore the Vulnerability Surface

Map where PJM's grid breaks. Move the sliders to explore the interaction between gas supply reliability and data center load. The white contour line marks the reliability threshold — everything above it means blackouts.

Gas Curtailment 0%

Fraction of gas generation lost during a cold snap. Elliott (2022) hit ~37%.

0%10%20%30%40%50%70%
Data Center Load 0GW

New always-on load from AI data centers. PJM queue: 40+ GW requested.

01530456075
Scenario Presets
Hours Unserved / Year
0
hours without power
RELIABLE
Max Deficit
0.0
GW
Unserved
0.0
TWh
People Affected
0
million
Vulnerability Surface — Hours Unserved
Gas Curtailment (%)
Hours unserved
Max deficit
Unserved energy
People affected
Data Center Load (GW)
Safe
0 hrs 8,760 hrs
Total failure
The Lesson

Two Risks, One Surface

Gas curtailment and data center load are not independent risks. They compound. A grid that survives 30% gas curtailment with no data centers fails catastrophically with 30 GW of new load added. The vulnerability surface reveals the nonlinear interaction that neither risk alone would expose.

Key Insight
The 9-hour reliability contour is not a straight line. It curves: each additional GW of data center load makes the grid more sensitive to gas curtailment. PJM cannot plan for either risk in isolation.

Try the Elliott Replay preset. During Winter Storm Elliott in 2022, PJM lost ~37% of gas generation. With 30 GW of data centers added, that same storm produces over 1,800 hours of unserved energy — and leaves tens of millions in the dark.

← Back PJM Data Center Study Next → Resilience Analysis