> **来源：[研报客](https://pc.yanbaoke.cn)** # Summary of "AI Infrastructure in the Age of Sovereignty: Requirements, Strategies and a Trusted Framework for Digital Embassies" ## Core Content This white paper explores the strategic importance of AI infrastructure in the context of growing national and geopolitical concerns, emphasizing the need for resilient and sovereign AI ecosystems. It outlines the evolving landscape of AI infrastructure, the strategies for building it, and the role of digital embassies in enabling international collaboration while maintaining control. ## Main Viewpoints - **AI Infrastructure as a Strategic Asset**: AI infrastructure is becoming a central determinant of national resilience and economic competitiveness, with global investments expected to exceed $400 billion annually by 2030. - **Sovereignty and Interdependence**: AI sovereignty is not about complete self-sufficiency but rather strategic interdependence. Economies must balance international collaboration with domestic control. - **Digital Embassies as a Solution**: Digital embassies are proposed as a means to extend AI infrastructure access beyond national borders, supported by a framework designed to enhance trust and security. - **Ecosystem Alignment**: For AI infrastructure strategies to succeed, policy-makers, investors, and other stakeholders must align on a clear strategic direction and future-ready system design. - **Resilience through Choices**: The paper highlights the importance of making informed, long-term infrastructure choices that consider both technical and institutional prerequisites. ## Key Information ### ① The AI Infrastructure Landscape AI infrastructure is composed of three core building blocks: - **Compute ("Data in Use")**: Refers to the processing power and capacity needed to train and deploy AI models. - **Connectivity ("Data in Motion")**: Describes the digital network infrastructure that connects data centres, edge resources, and end users. - **Data Storage ("Data at Rest")**: Involves secure systems and facilities for large-scale data management. Each building block can be configured in different ways: - **Compute**: Centralized, distributed, or endpoint-based. - **Connectivity**: Physical networks (e.g., fibre, subsea cables) or access networks (e.g., 5G/6G, satellite systems). - **Data Storage**: Centralized or distributed storage solutions, including cloud platforms and federated storage architectures. ### ② Designing AI Infrastructure Strategies Economies can choose from two reference strategies: 1. **Trusted International Partnerships**: Involves collaboration with reliable foreign partners, with shared governance and enforceable safeguards. 2. **Extensive Domestic Ownership**: Focuses on local control and development of AI infrastructure, often involving government or national platforms. Both strategies have their own set of **technical prerequisites** and **institutional prerequisites**: - **Technical Prerequisites**: - Energy: High demand, with mitigation strategies including renewable energy and distributed solutions. - Water: Critical for cooling, with solutions like recycled water and site selection based on water availability. - Land: Requires large areas with access to power and connectivity, with pre-zoning and streamlined approvals. - Hardware: Concentrated supply chains, necessitating diversification and local manufacturing incentives. - Cybersecurity: Layered security measures, real-time monitoring, and legal frameworks to protect data and systems. - **Institutional Prerequisites**: - Policy: Regulatory frameworks that support AI development and data governance. - Talent: Skilled workforce to manage and develop AI infrastructure. - Capital/Financing: Investment in infrastructure and long-term sustainability. ### ③ Creating Trust in Shared Arrangements: Digital Embassies - **Challenges**: Trust is a key issue in shared AI infrastructure, especially in cross-border collaborations. - **Benefits**: Trusted setups can enhance security, reduce dependency risks, and support continuity. - **Framework**: The **Global Framework for Innovative and Trusted Digital Embassies** is introduced to guide the development of resilient and secure digital embassies. - **Examples**: Estonia's use of a Luxembourg-based data storage facility under its control, and the European High-Performance Computing Joint Undertaking. ## Conclusion The paper calls for deliberate, strategy-led infrastructure choices to ensure long-term resilience and competitiveness. It emphasizes the need for **policy-makers to convene stakeholders** across public and private sectors to build and sustain sovereign AI ecosystems. The **Global Framework for Digital Embassies** is presented as a blueprint for secure and innovative cross-border AI infrastructure collaboration. ## Key Considerations for AI Ecosystem Actors - Align on a clear strategic direction and system design. - Jointly manage resource constraints (energy, water, land, hardware). - Enable trusted international partnerships. - Engineer resilience across cybersecurity, operations, and finance. - Leverage AI to monitor and protect critical infrastructure. This paper is a collaborative effort between the World Economic Forum and Bain & Company, contributing to the AI Global Alliance's work on AI competitiveness. It builds on previous insights and provides a practical foundation for making strategic infrastructure decisions in the context of AI sovereignty.