The solar corona is heated to ∼ 106 K over a chromosphere two orders of magnitude cooler, bymechanisms that are intrinsically intermittent and multi-scale: impulsive nanoflares from magneticreconnection and the turbulent dissipation of Alfvén waves. Resolving the microphysics (reconnec-tion at current sheets, wave cascades) over the hours-to-days evolution of a coronal loop imposesa computational barrier analogous to the Courant barrier in turbulence. We present a stochastic—subgrid framework that advances the loop energetics on macroscopic time steps while injecting theunresolved, intermittent heating as samples from a variance—gamma (VG) Lévy process, with asingle intermittency parameter ν controlling the rare, fat-tailed heating bursts. The non-monotonicoptically-thin radiative loss makes the loop thermally bistable, so a cool and a hot (coronal) branchcoexist; crossing to the hot branch is the analogue of a threshold "ignition." We show that with thesame mean heating, steady (Gaussian) heating cannot ignite the corona whereas VG intermittencydoes (ignition probability rising from 0% to ∼ 80% with ν): intermittency is a prerequisite for coronalheating. Extending the model to a one-dimensional loop with field-aligned conduction and chromo-spheric evaporation, and then deriving the heating rate and the intermittency self-consistently froma single magnetic-braiding profile Θ(x), we find that the corona lights up where braiding is high,with the active-region core ignited by mean heating (intermittency-independent) and the peripheryignited only by intermittency. Unifying nanoflares (DC) and Alfvén-wave turbulence (AC) as twoVG sources reproduces the observed thermal structure: waves supply a ∼ 1 MK floor everywherewhile nanoflares boost active regions to ∼ 2—3 MK, a spatial complementarity with distinct inter-mittency signatures. We compile seven predictions and present a verification methodology tied toexisting instruments (SDO, Solar Orbiter, Hinode, DKIST), including decision tables that fix inadvance which measured outcome refutes or robustifies the framework. A central, honest findingemerges: nanoflare heating tails are power-law with index α < 3 (divergent variance), exceeding thefinite-variance VG closure and calling for a stable-Lévy generalization, whereas wave heating fitsVG — the heating mechanism itself selects the closure