We study how streaming service fragmentation (SSF) can create media access friction (MAF), making illicit channels comparatively more attractive for some audiences. This friction is fueled by the current exclusivity model employed by the large streaming platforms. Decentralized, peer-to-peer (P2P) protocols like BitTorrent, through its trackerless DHT, present a robust illicit distribution channel. We introduce a cross-layer design that links the supply of torrents available via the BitTorrent DHT with the diffusion mechanics of specific torrent swarms around the release of new titles. Our macro panel (n=1,469 titles) shows that each additional legitimate provider of a given title reduces the instantaneous hazard of a first torrent by about 4.2%. In a targeted one-month micro panel of high-salience new releases, we tracked 244 torrent swarms involving 167,607 unique peer IPs across 220 jurisdictions and 7,277 ISPs. We find the torrents follow logistic S-curve diffusion with sharp and short-lived first mover advantage, while Bass imitation parameters are near zero, suggesting shock-driven growth rather than strong peer-to-peer contagion. Activity is highly concentrated by jurisdiction and network. Pre-release windows show elevated risk to users, with one-third of torrents carrying suspicious executables. Because BitTorrent operates as a federation of title-centric micro-networks with no stable chokepoint, broad interventions such as domain takedowns, CDN de-peering, or Layer-3 blocking do not map cleanly onto this distribution channel. The descriptive implication is that broader legitimate availability is a lower-collateral avenue for stakeholders to evaluate when seeking to reduce the demand conditions under which early leaks, early diffusion, and user cyber risk become most salient.