Turning idle smartphones into AI computing power: Where does the value of ACU actually come from? Recently, I've seen a lot of discussion about Acurast (@Acurast), especially @jiamigou who has a whole box of smartphones playing around with this, which piqued my interest to investigate its underlying concept. Basically, we can connect idle smartphones to form a decentralized computing network, using TEE technology to provide confidential and verifiable computing services for AI applications; In other words, Processor Lite can provide computing resources during the phone's idle time; Core is for qualified dedicated Android devices. Device operators can receive ACU rewards according to the rules. The barrier to entry seems relatively low, and overall it falls under the category of an AI × DePIN project. Regarding its token value logic: ACU computing fees generated by developers using their computing power are burned, and the network experiences a 5% annual token inflation to maintain token growth. Currently, Processor Lite can utilize idle mobile devices to provide computing resources, producing approximately 0.5 ACU per device per day; Core targets qualified dedicated Android devices, producing approximately 0.8 ACU per device per day. This also means that node rewards do not directly prove sufficient customer payments. It also depends on the number of continuously paying developers, actual computing expenditure, and the relationship between the burning amount and the issuance scale. The mainnet is now live, and ACU is available on Binance Alpha, Binance Perpetual Contracts, and Bithumb. In this era of AI computing power scarcity, how much computing power ACU can provide to AI through idle mobile phones worldwide is worth our continued attention. This is also why I believe Acurast deserves further observation.