After a brief introduction to NCAR, I'll discuss trends in climate modeling based on what we learned from simulations that supported IPCC. I'll argue that the strength and clarity of the IPCC AR4 conclusions are a direct result of increases in computing power that let us run higher resolution simulations, more ensembles, and better physics. Next steps will include regional analyses emphasizing hydrologic and biological cycles, coupled human-natural system models, and systems that aim to support decision-makers.