The rapid advancement of digital technologies and the growing complexity of financial systems have amplified the uncertainty, heterogeneity, and interconnectedness of modern financial problems. Traditional digital finance methods, which typically focus on single scenarios, isolated user groups, or narrowly defined tasks, struggle to address today’s hybrid financial environments characterized by cross-scenario demands, multi-stakeholder interactions, and tightly coupled functional and security requirements. To bridge this gap, we first propose a novel taxonomy of digital finance research encompassing five macro perspectives (Application, Model, Element, Platform Infrastructure, and Facility) providing a systematic coordinate system for positioning and analyzing existing studies. Building on this foundation, we introduce FAMEπ, a practical full-stack system architecture designed to empower secure, scalable, and holistic digital finance. FAMEπ co-designs and integrates diverse financial applications, a broad spectrum of models, theories of full-element aggregation optimization and trustworthy computation, and a unified intelligent computing platform. Unlike fragmented or patchwork solutions, FAMEπ achieves cross-layer interoperability and system-level optimization, enabling resilient, robust, and inherently secure digital finance across all major scenarios. This work thus establishes both a conceptual blueprint and an architectural path toward next-generation full-stack digital finance systems.
(102692) et al. (2026) studied this question.