The push to incorporate more renewable sources of generation into electricity grids is forcing a sea change in how grids are composed and managed. The operational task of balancing an electricity grid has grown from a relatively simple, solvable problem to one of great complexity, involving widespread prediction and coordination backed with substantial computational resources. New types of electric resources, particularly on the demand-side, are emerging in order to accommodate the fluctuations from this groundswell of renewable generation. In addition to the massive amount of capacity being added to utility electricity generation fleets, large amounts of distributed generation are being added on the customer side of electricity meters with unclear effects on grids as a whole. In this work, we assess the impact on the requirements for demand-side electricity resources from growth in both utility-scale and distributed renewable generation fleets. To do this, we construct models of the California electricity grid, both now and in a future scenario with 50% renewables penetration, as well as a model that incorporates data on over 125,000 rooftop photovoltaic installations. We then examine the context of these scenarios to understand the role and range of demand-side resources on these future, more renewable electricity grids. From this analysis, we show that a future grid for the state of California with 50% renewable generation and 5 times the distributed generation capacity of the current grid would reduce the overall fraction of total electricity delivered from thermal (natural gas) generation by over two-thirds, but would increase the frequency of large (±2GW) hourly ramps in thermal generation by a factor larger than five. Combating this enormous increase in ramps is a prime opportunity for demand-side management to increase the proportion of renewables as well as reduce the cost of electricity in future electric grids.