Recent quantum oscillation experiments on SmB6 pose a paradox, for while the angular dependence of the oscillation frequencies suggest a 3D bulk Fermi surface, SmB6 remains robustly insulating to very high magnetic fields. Moreover, a sudden low temperature upturn in the amplitude of the oscillations raises the possibility of quantum criticality. Here we discuss recently proposed mechanisms for this effect, contrasting bulk and surface scenarios. We argue that topological surface states permit us to reconcile the various data with bulk transport and spectroscopy measurements, interpreting the low temperature upturn in the quantum oscillation amplitudes as a result of surface Kondo breakdown and the high frequency oscillations as large topologically protected orbits around the X point. We discuss various predictions that can be used to test this theory.SmB 6 , discovered 50 years ago [1,2], has attracted recent interest due to its unusual surface transport properties: while its insulating gap develops around T K 50K, the resistivity saturates below a few Kelvin[3]. The renewed interest derives in part from from the possibility that SmB 6 is a topological Kondo insulator, developing topologically protected surface states at low temperatures [4][5][6][7]. Experiments [8][9][10] have confirmed that the plateau conductivity derives from surface states, and these states have been resolved by angle-resolved photoemission spectroscopy (ARPES) [11][12][13][14]. Furthermore, spin-ARPES experiments have revealed the spinmomentum locking of the surface quasiparticles expected from topologically protected Dirac cones [15].Yet despite this progress, some important experimental results are unresolved. In particular, quantum oscillation experiments on SmB 6 have given rise to two dramatically different interpretations [16,17]. Ref [16] observes low frequency (small Fermi surface) oscillations with the characteristic 1/ cos(φ) dependence on field orientation expected from 2D topological surface states. On the other hand Ref [17] detects a wide range of frequencies (both high and low frequency oscillations) which have been interpreted in terms of angularly isotropic three dimensional quasiparticle orbits, resembling a metallic hexaboride without a hybridization gap (such as LaB 6 ). A striking aspect of these measurements, is that the oscillations strongly deviate from a classic Lifshitz-Kosevich formula below ∼ 1K. Two recent theoretical proposals have been advanced to account for this bulk behavior, as a consequence of magnetic breakdown [18], of the formation of non-conducting Fermi surfaces [19].Another aspect of recent measurements, is the wide disparity in the reported effective masses of the carriers. The effective mass observed in both quantum oscillation experiments, m * /m ∼ 0.1 − 0.2 is an order of magnitude smaller than effective mass observed in ARPES [11][12][13][14] [17]. Error bars are the size of symbols which include both experimental errorbars and also errors from extracting data from a logarithmic plot. Given...