Fossil calibrations are used across vastly different phylogenetic scales, from constraining age estimates for the origin of diverse and ancient clades that have existed for 10s or 100s of millions of years, to far more recent divergences between closely related species. Often, fossil calibrations are used in a similar way regardless of these different scales. Here, I discuss how the information that fossils and phylogenies reflect about macroevolution changes across different scales, and I demonstrate the potential effects of this on divergence time estimates. Primarily, I show that at broader phylogenetic scales, fossil calibrations are far more likely to lead to the significant underestimation of node ages, even if the fossil preservation rate is constant. For such nodes, the discovery of plant fossils that are far older than previous node age estimates should be expected. Therefore, although the fossil record underpins much of our knowledge about the ages of different plant clades, the findings presented here demonstrate it is it is often the case that molecular data is incapable of extrapolating from known fossil ages to provide reliable node age estimates throughout a phylogeny. This reiterates the need to interpret node age estimates in a manner that is robust to these issues, and to develop models for lineage specific molecular rate variation that are less sensitive them.