The paper presents the features of applying the Nassim Taleb’s fragility theory to technical systems and structures. This theory allows analyzing stability and survivability of an object subjected to exceptionally large shock whose likelihood is close to impossible. For example, when an existing integral system (or a set of such systems), designed according to current standards, starts experiencing the impact of loads that have qualitatively changed their nature (for example, due to global climate change). The nature of catastrophes is associated with a strong interdependence of ongoing events. The concept of fragility can be used in risk analysis of technical systems for which loss distribution curves are considered, where losses are described by positive numbers and the right tail of the distribution is considered. Fragility, in the context of the problem under consideration, lies in the incorrect calculation of the risk from large-scale negative events. This is the so-called “simulation fragility”. That is, the systems are “fragile” to inaccuracies in estimating the distribution of stressors, and, consequently, to modeling errors, since this inaccuracy increases the likelihood of loads and impacts exceeding the design limits, leading to a greater likelihood of system failure. Fragility can be used to measure the non-linear response to a change in a model parameter by correlating fragility with model error.