While extreme precipitation intensity (EPI) is expected to increase with temperature due to the Clausius-Clapeyron relationship, previous studies at weather time scales indicate a clear upper bound to how much EPI can increase with warmer temperatures. After reaching a peak at a certain temperature (T_peak), EPI starts to decrease with temperature. This negative scaling at high temperatures has been attributed to the lack of moisture. To better understand the negative scaling, we investigate its possible dependence on storm types in the central US where extreme precipitation is produced by mesoscale convective systems (MCS) and other storm types (non-MCS) in the warm season. Hourly precipitation from observations is divided into one group associated with MCS and another group associated with non-MCS storms. By examining the EPI for each storm type, we find that EPI increases with temperatures that are lower than T_peak and plateaus within a ~10°C range before it decreases. This relationship applies to both MCS and non-MCS EPI, as reflected by two parallel EPI-T curves with the non-MCS curve featuring lower EPI than the MCS curve. The EPI-T curve using precipitation data from all storms lies between the two parallel curves but starts to decline at T_peak, shifting it closer to that of the non-MCS curve beyond T_peak. This happens because non-MCS events predominantly contribute to EPI at higher temperatures, while MCSs do not occur frequently at the higher temperature range. This is supported by further analysis of the diurnal cycle showing that MCS EPI primarily occurs in midnight hours with lower temperature and higher humidity, while non-MCS EPI mostly occurs in late-afternoon hours and can occur over a wide range of temperature and humidity. Therefore, our results suggest that a shift in storm type distribution over the diurnal cycle can contribute to the negative scaling of EPI-T, and analysis of the EPI-T scaling relationship for MCS and non-MCS precipitation separately may provide more insights on the EPI-T scaling and its implications for changes in EPI with future warming. The EPI-T relationship associated with MCS and non-MCS storms we find here will be used to evaluate how climate models capture such relationship with different storm types in the future.