1. Introduction: The 95% Problem In 2011, Google X (now X Development) proposed a radical solution to a persistent economic reality: while satellites offered global coverage but were expensive and high-latency, and cell towers offered high bandwidth but were geographically limited, nearly 95% of the world’s population lived within range of a cellular signal—yet only half were connected. The problem wasn't coverage; it was economic viability in rural and remote regions.
Loon’s envelope used helium. To lift a 15 kg payload (electronics + batteries) plus a 15 kg envelope, the balloon required displacing ~30 kg of air. At 20 km altitude (pressure ≈ 50 hPa), the volume needed is: google gravity balloon
Mathematically, the pressure differential (\Delta P) is limited by the meridional stress (\sigma) in the lobes: [ \Delta P = \frac{2 \sigma t}{R_{curv}} ] where (t) is film thickness and (R_{curv}) is lobe radius. By keeping (R_{curv}) small (many lobes), Loon could handle (\Delta P) up to 200 Pa without bursting. Unlike airships or drones, Loon had no propulsion. How do you steer a balloon? You change its altitude to catch different wind currents. The stratosphere has multiple layers of wind moving in different directions (e.g., west-to-east at 20 km, east-to-west at 25 km). Loon’s envelope used helium
Loon required —a fully sealed, rigid envelope that maintains internal pressure higher than the external atmosphere at all times. The challenge: as the sun heats the balloon, internal pressure rises, stressing the polyethylene film. By keeping (R_{curv}) small (many lobes), Loon could
That’s a sphere ~8.7 meters in diameter—roughly a tennis court’s width. The final Loon balloons used a pumpkin-shaped lenticular envelope to reduce drag and manage stress. Traditional weather balloons are zero-pressure : they have an open duct at the bottom. As gas expands (daytime heating, rising altitude), excess vents out. At night, the balloon contracts and descends. This is fine for a 2-hour radiosonde flight but disastrous for a 100-day mission.
The optimization problem: maximize the number of user-hours connected given constraints on battery (solar recharge rate), wind prediction error, and balloon longevity. This became a partially observable Markov decision process (POMDP) with >10^6 state variables.