As energy demand grows to power AI infrastructure, the grid expands into wildland. This is a simulation of what that means.
DESKTOP EXPERIENCE
This simulation uses cellular automata and interactive controls built for a large screen. Open on desktop for the full experience.
wired-to-burn.vercel.app
Designed by Pranavi Aourpally
Wired to burn
As energy demand grows to power AI infrastructure, the grid expands into wildland. More demand. More lines. More ignition points on drier ground. This is a simulation of what that means for the landscape we live in.
Wildfire terrain simulation
Six terrains. Three conditions. Your cursortap is a spark.
This simulation models how dryness, wind, and ignition points interact across different terrain types to spread wildfire. Each terrain has unique physics — spread rate, burn duration, moisture resistance. As energy demand increases, ignition points multiply. Lower moisture and higher wind accelerate everything. The math is real. The terrain is generated. What you see is what the physics produce.
1
Explore terrain markers to learn each type
2
Adjust terrain composition percentages
3
Change dryness, wind, and ignition below
4
Click or tap the landscape to place a spark
Terrain composition
0 burning · 0 burned · 0 alive
tap the landscape to place a spark
explore each terrain
increase dryness
add ignition points
Methodology
How this simulation works
This is a cellular automata wildfire simulation. The landscape is divided into a grid of cells. Each cell has a terrain type, a moisture level, and a state: alive, burning, or burned. Every few frames, the simulation checks each burning cell and calculates whether the fire spreads to its neighbors.
The spread formula
For each neighbor of a burning cell, the chance of catching fire is:
spread_chance = spread_rate × (1 − moisture) if moisture ≥ ignition_threshold → chance × 0.05 if wind active → chance × (1 + dot_product × wind_speed × 4)
spread_rate is how flammable the terrain is. Moisture is controlled by the dryness slider. The dot product calculates wind direction influence — downwind cells catch fire faster, upwind cells are protected.
The six terrain types
Terrain
Spread
Burn
Moisture
Ignites
Recovery
Forest
4%
28f
70%
<50%
0.1%
Shrubland
8%
22f
35%
<45%
0.3%
Agriculture
6%
16f
50%
<40%
0.8%
Grassland
12%
12f
45%
<35%
1.2%
Wetland
1%
35f
90%
<25%
0.6%
Developed
2.5%
20f
55%
<40%
0%
What the controls do
Dryness reduces moisture. At high dryness, even wetland loses fire resistance.
Wind multiplies spread in its direction. High wind = fire races one way.
Ignition points are potential spark sources — more energy demand means more risk.
Terrain composition changes the landscape mix via the picker.
Designed and developed by Pranavi Aourpally
Creative technologist · UI/UX designer
Presented at TIAT (The Intersection of Art & Technology), San Francisco, 2026
Sources: USFS fire behavior models · IEA Electricity 2024 · NIFC · NLCD 2016 · p5.js + Cellular automata This is a simulation model showing contributing factors — not a prediction.
Wired to burn
drier air, faster ignition
70%
pushes fire, feeds oxygen
60%
N
more demand, more risk
20
regenerate
how this works
a simulation model — not a prediction · designed by Pranavi Aourpally · presented at TIAT SF 2026