In order to realise scalability of chemical sensors in extensively deployed wireless sensor networks, considerable materials challenges must be overcome. Conventional devices are currently far too expensive and unreliable for massive long-term field deployment. Cost can be driven down by imaginative approaches to transduction and instrument design. For example, we have produced a complete instrument based on LED measurement of colour changes that has sub-micromolar detection limits for a number of heavy metals for around $1.2 In its current form, the device also has a short distance wireless communications functionality and very low power consumption. However, chemical sensors capable of long-term reliability will require imaginative solutions to the key issue -how can the sensing films/membranes in chemical sensors maintain predictable characteristics in long term deployment?The vision of 'internet-scale sensing' will only be realised through advances in materials science and a complete rethink of how we do chemical sensing. For example, fully autonomous sensing platforms must be completely self-reliant in terms of power, communications, reagents and consumables. The sensor network must be self-sustaining, meaning that as individual nodes become unreliable, new nodes are established, for example through physical replacement or through devices capable of self-repair/regeneration. In this chapter, these issues are presented, along with some recent advances mentioned above.