In the near future, most IoT devices will communicate with each other instead of powerful access points. It is therefore increasingly important to finnd end-to-end, robust and scalable security solutions for these devices, especially since many of these devices communicate very sensitive information, such as health-related data. Given the tight size and energy constraints of IoT sensors, designing such security solutions is a hard problem. We believe software-only security will not be sufficient to overcome these constraints, and that future IoT technologies require complementary software and hardware security solutions. We are currently exploring software-hardware techniques to build IoT systems which can extract the location of a sensor from it's RF signal at the physical layer, and pass its communications to the higher layers only if the location is valid. A system which can detect these finngerprints would be a very powerful security mechanism, capable of reliably identifying counterfeit nodes.