2026-05-11
Every storm season the same comment surfaces online: "Why don't we just harvest lightning?" Let's actually run the numbers and see how badly the universe punishes this idea.
Energy per strike. A typical cloud-to-ground bolt dissipates around 1 GJ (1 billion joules), with monsters reaching ~5 GJ. One gigajoule equals 278 kWh — about 9 days of average US household electricity. Sounds promising. Then reality intervenes.
The peak-power problem. That energy is delivered in roughly 30 microseconds. Power = energy/time = 10⁹ J ÷ 3×10⁻⁵ s ≈ 33 TW. For a 30 μs window, you must absorb twenty times the entire planet's electrical generating capacity. No wire, no transformer, and certainly no battery survives that. You need a capacitor.
Sizing the capacitor. Energy in a cap is E = ½CV². At a return-stroke voltage of 100 MV:
C = 2E/V² = 2(10⁹) / (10⁸)² = 2×10⁻⁷ F = 0.2 μF
The capacitance itself is trivial — a coin-sized ceramic. The killer is the voltage rating. Commercial high-voltage caps top out near 100 kV. We need one rated 1000× higher, with dielectric breakdown strength beyond anything in the materials database. Stacking 1,000 caps in series? Now you need perfectly matched leakage and a switch that closes in nanoseconds at 100 MV. Spark gaps work, but they are the lossy part of lightning.
Where does the energy actually go? Of that 1 GJ, roughly:
So our optimistic harvest is 14 kWh per strike — half a day of household power. After capacitor losses, switching losses, and DC-DC conversion from 100 MV down to 240 V (efficiency maybe 30%), call it 4 kWh per bolt.
Scaling up: the global cap. Earth sees ~44 flashes/second, or 1.4 billion strikes/year. Catching every single one (impossible — most hit ocean or unwired ground) yields:
1.4×10⁹ × 4 kWh = 5.6 TWh/year
Global electricity demand is ~30,000 TWh/year. We've harvested 0.019% — about what one mid-sized natural gas plant produces. To do that, we have wired every square kilometer of the planet with kilometer-tall lightning rods.
Best-case site: Catatumbo, Venezuela. The famous "lightning capital" sees ~250 storm nights/year and ~1.2 million strikes/year over a small lake. A 300 m steel tower with a 1 km strike radius might intercept ~50 bolts/year. Output: 200 kWh/year. A single rooftop solar panel beats it by a factor of 2.
The deeper irony. The Earth itself is already a giant lightning-powered capacitor. The fair-weather atmospheric circuit carries ~1,800 A globally at ~300 kV — roughly 500 MW sustained, fed continuously by thunderstorms. Tapping that circuit (Tesla's Wardenclyffe dream) at least gives you a steady-state source. But it's still less power than a single coal unit, distributed over the entire planet.
Lightning is breathtaking, but as an energy resource it fails three ways at once: too peaky to capture, too rare to scale, and too leaky to transmit.