It is forecast that there will be 62 billion connected devices by 2024 rising to 350 billion by 2030. This growth will be driven by a sensor enabled metaverse, the internet of things and edge computing along with industry initiatives such as tiny machine learning (tinyML). The usual definition of tinyML is running machine learning on embedded devices at an average of less than one milliwatt in power. This power requirement is important because it allows unattended devices running on batteries or energy harvesting. If we postulate that half of the 62 billion connected devices will be battery powered, and that of these batteries’ half will need to be replaced every year, this equates to 60M battery replacements needed per working day, which is infeasible.