The Universe is opaque to extragalactic very high-energy gamma rays (VHEGRs, E > 100 GeV) because they annihilate and pair produce on the extragalactic background light. The resulting ultra-relativistic pairs are commonly assumed to lose energy primarily through inverse Compton scattering of cosmic microwave background photons, reprocessing the original emission from TeV to GeV energies. In Broderick et al. (2012, Paper I of this three paper series), we argued that this is not the case; powerful plasma instabilities driven by the highly anisotropic nature of the ultra-relativistic pair distribution provide a plausible way to dissipate the kinetic energy of the TeV-generated pairs locally, heating the intergalactic medium (IGM). Here, we explore the effect of this heating upon the thermal history of the IGM. We collate the observed extragalactic VHEGR sources to determine a local VHEGR heating rate. Given the pointed nature of VHEGR observations, we estimate the correction for the various selection effects using Fermi observations of high and intermediate peaked BL Lacs. As the extragalactic component of the local VHEGR flux is dominated by TeV blazars, we then estimate the evolution of the TeV blazar luminosity density by tying it to the well-observed quasar luminosity density, and producing a VHEGR heating rate as a function of redshift. This heating is relatively homogeneous for z 4, but there is greater spatial variation at higher redshift (order unity at z ∼ 6) because of the reduced number of blazars that contribute to local heating. We show that this new heating process dominates photoheating in the low-redshift evolution of the IGM and calculate the effect of this heating in a one-zone model. As a consequence, the inclusion of TeV blazar heating qualitatively and quantitatively changes the structure and history of the IGM. Due to the homogeneous nature of the extragalactic background light, TeV blazars produce a uniform volumetric heating rate. This heating is sufficient to increase the temperature of the mean density IGM by nearly an order of magnitude, and at low densities by substantially more. It also naturally produces the inverted temperature-density relation inferred by recent observations of the high-redshift Lyα forest, a feature that is difficult to reconcile with standard reionization models. Finally, we close with a discussion on the possibility of detecting this hot low-density IGM suggested by our model either directly or indirectly via the local Lyα forest, the Comptonized cosmic microwave background, or free-free emission, but find that such measurements are currently not feasible.