• Studies of the macroevolutionary legacy of paleopolyploidy are limited by an incomplete sampling of these events across the tree of life. To better locate and understand these events, we need comprehensive taxonomic sampling as well as homology inference methods that accurately reconstruct the frequency and location of gene duplications.</p><p>• We assembled a dataset of transcriptomes and genomes from 169 species in Caryophyllales, of which 44 were newly generated for this study, representing one of the densest sampled genomic-scale datasets yet available. We carried out phylogenomic analyses using a modified phylome strategy to reconstruct the species tree. We mapped phylogenetic distribution of paleopolyploidy events by both tree-based and distance-based methods, and explicitly tested scenarios for paleo-allopolyploidy.</p><p>• We identified twenty-six paleopolyploidy events distributed throughout Caryophyllales, and using novel techniques inferred two to be paleo-allopolyploidy.</p><p>• Through dense phylogenomic sampling, we show the propensity of paleo-polyploidy in the clade Caryophyllales. We also provide the first method for utilizing transcriptome data to detect paleo-allopolyploidy, which is important as it may have different macro-evolutionary implications compared to paleo-autopolyploidy.