Physically, the air in front of the meteor is experiencing adiabatic compression. The jump values of density and temperature before and after the adiabatic shock are given by the Rankine-Hugoniot conditions.
The analysis of those conditions reveal that the density jump is limited to a finite value, but the temperature jump before and after the shock is proportional to the square of the shock (i.e. the meteors) mach number $\Delta T\sim M^2$ (source: Mihalas&Mihalas, Foundations of Radiation Hydrodynamics, Section 104, "Steady shocks"), hence in principle unlimited.
This gives rise to the enormous temperatures experienced by meteors. Of course there are multiple complications to this picture, due to ionization, turbulence, boundary layer formation etc. but this is the starting point for a physically correct reasoning.
The myth about air friction somehow became common knowledge, but essentially the viscosity of air doesn't play that much of a role. While the value of viscosity is important for the heat transport via conduction, this process is however negligible, as radiative and turbulent energy transport dominate in the domain around the meteor.
Finally, the friction or viscosity does play an important theoretical role: In general, the pre-and post-shock values of hydrodynamic values given by the Rankine-Hugoniot conditions can have different entropies. Regions of different entropies must be connected by some dissipative mechanism, which can be given by an enormously thin viscous layer in the shock structure. However this hardly merits the statement that "air friction heats meteors".