Future large ensembles of time delay lenses have the potential to provide interesting cosmological constraints complementary to those of other methods . In a flat universe with constant { w } including a Planck prior , LSST time delay measurements for \sim 4 , 000 lenses should constrain the local Hubble constant h to \sim 0.007 ( \sim 1 \% ) , \Omega _ { de } to \sim 0.005 , and { w } to \sim 0.026 ( all 1- \sigma precisions ) . Similar constraints could be obtained by a dedicated gravitational lens observatory ( OMEGA ) which would obtain precise time delay and mass model measurements for \sim 100 well-studied lenses . We compare these constraints ( as well as those for a more general cosmology ) to the “ optimistic Stage IV ” constraints expected from weak lensing , supernovae , baryon acoustic oscillations , and cluster counts , as calculated by the Dark Energy Task Force . Time delays yield a modest constraint on a time-varying { w } ( z ) , with the best constraint on { w } ( z ) at the “ pivot redshift ” of z \approx 0.31 . Our Fisher matrix calculation is provided to allow time delay constraints to be easily compared to and combined with constraints from other experiments . We also show how cosmological constraining power varies as a function of numbers of lenses , lens model uncertainty , time delay precision , redshift precision , and the ratio of four-image to two-image lenses .