📄 中文摘要
人工智能主权是指国家、组织或个人能够根据自身的法律、价值观和安全需求，自主开发、运行和控制人工智能系统，而不完全依赖于外国或不透明的提供者。自托管大型语言模型（LLM）可以确保数据、模型和推理过程的控制权，为团队、企业和国家实现人工智能主权提供了一条切实可行的路径。文章探讨了构建主权人工智能的各个方面和方法，以及各国在应对这一挑战时的策略和措施。
大型语言模型自托管与人工智能主权
出处: LLM Self-Hosting and AI Sovereignty
发布: 2026年2月22日
📄 English Summary
LLM Self-Hosting and AI Sovereignty
AI sovereignty refers to the ability of a country, organization, or individual to develop, operate, and control AI systems according to their own laws, values, and security needs, rather than relying entirely on foreign or opaque providers. Self-hosting large language models (LLMs) allows for the control of data, models, and inference processes, providing a practical path toward AI sovereignty for teams, enterprises, and nations. The article explores various aspects and methods used to build sovereign AI, as well as how countries are addressing the challenges associated with it.