There is concern that some games obtain most of their revenue from a small proportion of heavily involved individuals; that some games may systematically encourage heavy spending amongst this group; and that this expenditure may be financially burdensome. We explore these questions using a transactional dataset of USD 4.7bn in in-game spending drawn from 69,144,363 players of 2,873 mobile games over the course of 624 days. Distributional cluster analysis shows that a subset of the mobile games market (‘Hyper-Pareto games’) make a large proportion of their profits (~38%) from 1% of their spenders (n=338, 11.8% of games in sample). Cross-game comparisons reveal that a typical game in this cluster has its top 1% spend USD 1710.60; and that the more a game relies on its top 1% for revenue generation, the more these individuals tend to spend. We find that specific genres of game are typified by higher spending, with simulated gambling products having the highest levels of in-game spending amongst any genre: For example, a typical family genre game has an average spend of USD 30.62 amongst its top 1%; by contrast, a casino game, USD 2,381. We find evidence of a high degree of variability in the amount spent by the top 1% between games: In a small subset of games the top 1% of gamers spend tens of thousands of dollars. We discuss the implications for future studies on links between gaming and wellbeing.