At 23:24 UTC on 29 July 2025, the seafloor 119 km off Petropavlovsk-Kamchatsky ruptured in an Mw 8.8 megathrust, the most powerful earthquake anywhere since Tohoku 2011 and joint-fourth-largest of the last half-century. The rupture unzipped roughly 600 km of the Kuril-Kamchatka subduction interface, re-breaking the source patch of the great 1952 event; a Mw 7.4 foreshock nine days earlier was, in hindsight, a warning shot. Yet the human toll barely registered: no direct deaths, one indirect death in Japan, about 25 injuries and roughly 1,400 damaged homes. A Pacific-wide tsunami followed, running 3 to 6 metres near the Kamchatka and Kuril coast but under two metres across the ocean, and prompting mass precautionary evacuations of around two million people in Japan and 1.5 million in Chile.
The outcome is the mirror image of Kahramanmaras: the energy was enormous, but there was almost nothing in the way. The epicentre was offshore and deep on the interface, the nearest sizeable town sits 120 km away, the peninsula is among the most sparsely populated places on Earth, and a tsunami-warning system that fired within ten minutes carried lives toward zero. For a risk function the lesson is twofold. The small loss that did occur was in proportional terms, and that keeps the international market out altogether, so a sanctioned territory is structurally uninsurable from the outside. The same subduction machine pointed at Cascadia, Nankai or central Chile would be a capital event.