This proposal aims to study an advanced computing platform for the execution of privacy-sensitive sensing tasks in integrated communications and sensing (ISAC) Wi-Fi networks. Private machine learning (ML)-based sensing tasks can be (i) pre-processed on access points (APs) and routers, (ii) offloaded to edge computers co-powered by renewables with small GPUs, (iii) offloaded to the Amazon Web Services (AWS) cloud, or (iv) a combination of the previous approaches. By optimizing offloading strategies, we target the best tradeoff among energy efficiency, latency, and privacy protection in a scalable and environmentally responsible manner. This will ultimately enable a privacy-preserving sensing system through artificial intelligence (AI) that is efficient and locally carbon neutral.