The increase in connected device, linked with the evolution of IoT and cellular networks with 5G and towards 6G makes crucial the limitation of the interference impact. Devices in networks (like LoRa, SigFox and even WiFi) or in different networks are not necessarily coordinated to optimize their communication scheme and reduce interference. As a consequence, it is essential to characterize the interference statistics in order to assess its impact on the network performance and define the appropriate access policies. With a theoretical approach, a large number of studies have tackled this characterization question. However, much less is available from an experimental point of view. In this paper, we address this key gap. Measurements were performed in Aalborg, Denmark, in the unlicensed 863–870 MHz. We show that the heavy tailed behaviour of the interference, predicted by theory, is indeed appropriate. We also show that the α-stable distribution can be a good model for interference in IoT networks.