2026主权时代的AI基础设施白皮书：数字大使馆建设需求、发展战略与可信框架

AI Infrastructure inthe Age of Sovereignty:Requirements, Strategies W H I T EP A P E R Contents Foreword Executive summary 1The evolving artificial intelligence infrastructure landscape 1.1Dynamics shaping artificial intelligence infrastructure1.2The foundation of AI infrastructure 2Designing AI infrastructure strategies 2.1Defining the AI sovereignty spectrum and its reference strategies 92.2Examining local requirements of reference strategies 3Expanding AI infrastructure through digital embassies 3.1Key challenges to overcome to realize digital embassies3.2Key benefits of trusted setups Conclusion: Key considerations for AI ecosystem actors Appendix Contributors Endnotes Disclaimer This document is published by theWorld Economic Forum as a contributionto a project, insight area or interaction.The findings, interpretations andconclusions expressed herein are a resultof a collaborative process facilitated and ©2026 World Economic Forum. All rightsreserved. No part of this publication maybe reproduced or transmitted in any formor by any means, including photocopying Foreword Cathy LiHead, Centre for AIExcellence; Member,Executive Committee,World Economic Forum Florian MuellerSenior Partner and Head,AI, Insights & Solutions forEurope, Middle East and Artificial intelligence (AI) infrastructure is rapidlybecoming one of the most consequential strategicassets of the digital age. What was once viewed as binding resource constraints – from energy andland to hardware availability – are reshaping theAI infrastructure landscape. Decisions madetoday will shape an economy’s AI infrastructure Cumulative investment in AI-dedicatedinfrastructure exceeded $600 billion between 2010and 2024 and is projected to surpass $400 billionannually by 2030.1These investments matter forreasons that extend well beyond technologicaladvancements. Growing reliance on increasinglycomplex financing structures and greater use ofdebt mean that disruptions or underperformancecan have spillover effects across the wider financial This paper, in collaboration with Bain & Company,is part of the World Economic Forum’s AI GlobalAlliance’s work on AI competitiveness. It buildson the Forum’s previous thought leadership in this series:Blueprint for Intelligent Economies:AI Competitiveness through Regional CollaborationandRethinking AI Sovereignty: Pathways toCompetitiveness through Strategic Investments.This paper describes AI infrastructure as a core system. At the same time, AI infrastructure isbecoming increasingly integral to the deliveryof public services. Out of 33 Organisation forEconomic Co-operation and Development (OECD) Notably, the paper explores digital embassiesas one option to extend access to AI infrastructurebeyond national borders. Trust is a central challengein shared AI infrastructure. The Forum and a globalcommunity of stakeholders have co-designed aframework to reduce uncertainty and strengthen Against this backdrop, the motivation to pursueAI sovereignty is stronger than ever. Yet formost economies, self-sufficiency is unrealisticgiven the scale, complexity and concentrationof AI infrastructure required. Therefore, AIsovereignty must be pursued through strategicinterdependence – built through deliberate choicesabout where to rely on trusted international Above all else, this paper is a call to action foreconomies to make deliberate, strategy-ledinfrastructure choices. We encourage policy-makersto convene AI ecosystem stakeholders across the Executive summary Artificial intelligence sovereignty ambitionsare rising alongside the stakes. Today’s Artificial intelligence (AI) infrastructure3has movedto the centre of global debates on competitiveness.Geopolitical tensions, the digitization of publicservices and intricate financing structures havemade AI infrastructure a strategic concern –rather than a purely technological one. As demandfor compute and data storage accelerates and economies can select a strategic direction (i.e.higher or lower interdependence) and define their AIinfrastructure strategy in line with local capabilities. To inform this decision-making, this paperdescribes two reference strategies at opposite 1.Trusted international partnerships–where sovereignty is achieved through 2.Extensive domestic ownership–anchored in local control of AI infrastructure The AI infrastructure For each reference strategy, this paper outlinesthe AI infrastructure building blocks and technicalprerequisites that must be locally available for thestrategy to succeed. These reference strategies –trusted international partnerships and extensivedomestic ownership – are not binary choices. The AI infrastructure landscape comprises threebuilding blocks, each corresponding to a different Compute(“data in use”) refers to computingpower and processing capacity used to trainand deploy AI models and applications. Creating trust in sharedarrangements:a framework Connectivity(“data in motion”) describesthe digi