This paper proposes a concept to construct a community-based cross-industrial data/contents delivery plat-form named “Piggy-back network which utilizes already-existing local fixed and mobile resources in a smart city. These fixed and mobile resources are assumed to equip store-carry-forwarding-based (SCF-based) wireless data sharing functions, i.e., a short-range but extremely high-speed millimeter-wave device, a mid/long-range but low-speed microwave device, and data storage butter. It is discussed that the data delivery performance of such SCF-based platform could exceed the one when using wired/wireless infrastructure directly, and it will be significantly improved if an AI-based mobility prediction engine recommends the human drivers or the driving controllers of future automated driving vehicles to detour and/or stop by some specific locations. It is theoretically shown that the mobile resources can potentially deliver high-volume data with shorter time than using wired/wireless network infrastructures under some conditions. The real commercial mobilities' trajectory data obtained experimentally in the city of Kakogawa, Japan, are analyzed and the potential of the data delivery performance is estimated.