How you feed and supply fifty-two thousand displaced people when the roads are broken and the bridges are cracked — and the trucks were already rolling before the ground shook.
Traditional disaster logistics is reactive: the disaster hits, someone requests aid, warehouses hundreds of miles away start loading. By the time it arrives, the first seventy-two hours — the ones that decide who lives — are gone. Carlos Mendoza, directing the PHIN0 logistics AI, ran the opposite play. On September 30, with no confirmed disaster, three convoys — thirteen tractor-trailers of medical supplies, food, water, shelter, and comms — were pre-positioned at staging points a hundred miles out, around Jackson, Tennessee.
Because they were already moving, the first supplies reached Memphis six hours and fifteen minutes after the quake — sixty-six hours ahead of a reactive timeline. PHIN0 didn’t decide to spend the money; Commander Ortega did. PHIN0 told her exactly what to load, where to stage it, and which route would still be standing.
Supplying a disaster isn’t about how much you have — it’s about matching what you move to what’s actually needed, block by block, hour by hour, on infrastructure that keeps failing. Faith charities coordinated through THE NET delivered 22,000 meals a day during recovery. Steve Erkal’s overbuilt warehouse became a distribution node as well as a shelter for five hundred.
And the honest part — the after-action found food demand was underestimated by 34%. That miss became a training module. The system worked; it also learned exactly where it nearly didn’t. That’s the whole doctrine: count the save and the gap with the same pen.