Searches for the characteristic large-angle B-mode polarization pattern in the CMB, which would be produced by a stochastic background of primordial gravitational waves, have so far yielded only upper limits on the tensor-to-scalar ratio (e.g. r ? 0.03–0.04 at 95% confidence), with earlier claimed detections by BICEP2 shown to be dominated by polarized dust foregrounds (BICEP2/Keck Collaboration 2015; Planck Collaboration XI 2016). In the minimal ΛCDM plus single-field slow-roll inflation framework, many theoretically attractive inflation models naturally predict tensor amplitudes that are now disfavored, forcing either fine-tuned inflationary potentials, modified initial states, or more complex scenarios to reconcile the absence of detected primordial B-modes with expectations from simple high-energy inflation (Lyth 1997; Martin et al. 2014).