Abundant evidence show that “stable” cratons are modified during their existence, as illustrated in several cases: The Baltic Shield formed during the Svecofennian Orogeny around 1.7 Ga and its western parts were reworked by the Sveconorwegian/Grenvillian orogeny. Recent geophysical interpretations image a large body of crustal material in eclogite facies beneath the present Moho in the central shield. This body probably formed after the initial cratonisation (Buntin et al., Nature Comm. 2021). The isopycnicity hypothesis proposes that a trade-off between composition and temperature of the lithospheric mantle maintains constant topography in cratons (Jordan, Nature 1978) based on kimberlite data from South Africa. However, gravity data from Siberia shows that kimberlite pipes solely modify cratons in isostatic equilibrium (Artemieva et al., EPSL 2019). Therefore, kimberlite sampling is non-representative, and the real composition of most cratonic mantle lithosphere is unknown. Strong seismic anisotropy is observed in many cratons and is commonly attributed to the mantle due to frozen-in lithospheric features or asthenospheric flow. Recently it was demonstrated that a major part of the anisotropy resides in the crust of the Kalahara craton and that the fast axes are parallel to the strike of major dyke swarms and orogenic fabric (Thybo et al., Nature Comm. 2019). This finding indicates significant craton modification by magmatic intrusion. Modification by external stresses and induced magmatism even may split existing cratons. Integrated interpretation of existing data and geodynamic modelling show that a linear sequence of volcanic harrats in the Arabian craton potentially represents the formation of a new plate boundary (Artemieva et al., Earth Science Review 2022). It is probable that the extension in the northern Red Sea rift will jump to the volcanic lineament, which eventually will develop into new ocean spreading and effectively split the existing craton.