With polymer nanocomposites, achieving highly effective dispersion of agglomerated nanofiller and major or optimal property enhancements remain challenges. A commonly posited solution is to improve the polymer-filler surface thermodynamic compatibility; this approach has led to significant improvements in some cases, but it has not provided a general solution. We address the question of whether achieving a metastable, well-dispersed state is better than compatibilization in attaining the goal of major property enhancements. We use solid-state shear pulverization to produce well-dispersed polypropylene (PP) nanocomposites with up to 8 or 9 wt% pristine nanosilica (p-NS) or organically modified nanosilica (m-NS). Microscopy shows that as-received, tens-of-micron-sized p-NS and m-NS agglomerates undergo very good dispersion, with similar to 10-100 nm size-range nanofiller in hybrids. Rheology is consistent with very good dispersion, with only 92/8 wt% PP/p-NS indicating incipient nanofiller network formation. The PP/p-NS hybrids have superior Young's modulus and tensile strength. Relative to PP, modulus increases by 22% and 12% and tensile strength by 19% and 14% for 99/1 wt% PP/p-NS and 99/1 wt % PP/m-NS, respectively. The PP/p-NS hybrids have the largest increases in modulus (46% at 8 wt% p-NS) and tensile strength (22% at 6 wt% p-NS). Upon melting and crystallization, both PP/p-NS and PP/m-NS result in PP beta-crystal formation at 1 wt% nanosilica, with p-NS having a greater effect. The PP/p-NS hybrid shows larger increases in thermal stability and nucleating efficiency for PP crystallization. Thus, with very good dispersion, unmodified nanofiller in a metastable dispersed state can result in more robust nanocomposites than when modified nanofiller is used to compatibilize the polymer nanofiller interface. (C) 2015 Elsevier Ltd. All rights reserved.