Gravity and most seismic interpretations agree that the Sierra Nevada is at present isostatically compensated by a crustal root. The most reasonable timing for emplacement of the root was during Mesozoic batholithic intrusions. This is difficult to reconcile with the evidence that major uplift of the mountains occurred in the past 10 m.y. A simple quantitative model for flexural isostasy of an elastic plate before and after breaking resolves this problem and explains the tilt of the Sierra Nevada block and variations in topography along the range. We postulate that cooling and elastic strengthening of the Mesozoic magmatic arc prevented its reaching local isostatic equilibrium during erosion in the Cenozoic. Thus, an overcompensated residual mountain range was held down elastically until Basin and Range extension broke the elastic plate along the Owens Valley and allowed rapid uplift. The Moho density contrast that best models the amplitude of both late Cenozoic uplift and prior topography is slightly larger than seismic and gravity models suggest; this implies that up to 20% of the uplift may be driven by density contrasts in the upper mantle. The idea of overcompensated erosional remnants and their release upon lithospheric faulting has a more general application to “anorogenic” uplift of mountain blocks.