This paper presents the first non-perturbative study of single-field inflation with enhanced density fluctuations, focusing on the impact of small-scale features in the inflationary potential. The research shows that oscillatory features in the potential can drastically alter the course of inflation, leading to significant phenomenological consequences. In some cases, the entire Universe gets trapped in a forever inflating de Sitter state, while in others, only some regions get stuck in a false vacuum, offering an alternative channel for primordial black hole (PBH) formation. These results highlight the "inflationary butterfly effect," where small-scale phenomena can have profound consequences on the evolution of the entire Universe. The study uses lattice field theory simulations to explore the small-scale physics of inflation, particularly in the regime relevant for gravitational-wave astronomy. The simulations reveal that tiny quantum fluctuations, amplified by the oscillatory feature, can drastically affect the fate of the entire Universe. This underscores the need for non-perturbative methods to understand the small-scale physics of inflation and to make firm predictions for testing against observations. The paper discusses three cases: a perturbative case, a highly non-perturbative case, and a non-perturbative case. In the perturbative case, the inflaton field behaves as expected, with small fluctuations. In the highly non-perturbative case, the inflaton gets trapped in the oscillatory part of the potential, leading to a de Sitter state. In the non-perturbative case, only some patches of the Universe are trapped in the oscillatory region, offering an alternative channel for PBH formation. The study also shows that the fraction of the volume in the simulation box in which the inflaton is trapped at the end of the simulation, F, is extremely sensitive to variations in the parameters. The results indicate that to avoid PBH overproduction, the parameter range should be such that F is less than 10^{-6}. The paper also discusses the implications of these findings for gravitational-wave astronomy, highlighting the potential of lattice simulations to provide insights into these phenomena.This paper presents the first non-perturbative study of single-field inflation with enhanced density fluctuations, focusing on the impact of small-scale features in the inflationary potential. The research shows that oscillatory features in the potential can drastically alter the course of inflation, leading to significant phenomenological consequences. In some cases, the entire Universe gets trapped in a forever inflating de Sitter state, while in others, only some regions get stuck in a false vacuum, offering an alternative channel for primordial black hole (PBH) formation. These results highlight the "inflationary butterfly effect," where small-scale phenomena can have profound consequences on the evolution of the entire Universe. The study uses lattice field theory simulations to explore the small-scale physics of inflation, particularly in the regime relevant for gravitational-wave astronomy. The simulations reveal that tiny quantum fluctuations, amplified by the oscillatory feature, can drastically affect the fate of the entire Universe. This underscores the need for non-perturbative methods to understand the small-scale physics of inflation and to make firm predictions for testing against observations. The paper discusses three cases: a perturbative case, a highly non-perturbative case, and a non-perturbative case. In the perturbative case, the inflaton field behaves as expected, with small fluctuations. In the highly non-perturbative case, the inflaton gets trapped in the oscillatory part of the potential, leading to a de Sitter state. In the non-perturbative case, only some patches of the Universe are trapped in the oscillatory region, offering an alternative channel for PBH formation. The study also shows that the fraction of the volume in the simulation box in which the inflaton is trapped at the end of the simulation, F, is extremely sensitive to variations in the parameters. The results indicate that to avoid PBH overproduction, the parameter range should be such that F is less than 10^{-6}. The paper also discusses the implications of these findings for gravitational-wave astronomy, highlighting the potential of lattice simulations to provide insights into these phenomena.