Nanocomposite materials may be considered as a low-risk application of nanotechnology, if the nanofillers remain embedded throughout the life-cycle of the products in which they are embedded. We hypothesize that release of free CNTs occurs by a combination of mechanical stress and chemical degradation of the polymer matrix. We experimentally address limiting cases: Mechanically released fragments may show tubular protrusions on their surface. Here we identify these protrusions unambiguously as naked CNTs by chemically resolved microscopy and a suitable preparation protocol. By size-selective quantification of fragments we establish as a lower limit that at least 95 % of the CNTs remain embedded. Contrary to classical fiber composite approaches, we link this phenomenon to matrix materials with only a few percent elongation at break, predicting which materials should still cover their CNT nanofillers after machining. Protruding networks of CNTs remain after photochemical degradation of the matrix, and we show that it takes the worst case combinations of weathering plus high-shear wear to release free CNTs in the order of mg/m 2/year. Synergy of chemical degradation and mechanical energy input is identified as the priority scenario of CNT release, but its lab simulation by combined methods is still far from real-world validation.