Sovereign AI

Sovereign AI is a nation's capability to independently develop, train, and operate artificial intelligence models on domestically located, and often domestically owned or controlled, compute infrastructure, data, and talent, without primary dependency on foreign hyperscalers. The term was popularized by NVIDIA CEO Jensen Huang from 2024 onward as governments launched national AI compute initiatives: Japan funded domestic GPU clusters through METI, France backed Mistral AI and national compute capacity, Saudi Arabia formed Humain, and the UAE built out G42's sovereign cloud alongside Microsoft and OpenAI partnerships. In Southeast Asia, sovereign AI concerns increasingly shape data center site selection and licensing, with governments in Malaysia, Indonesia, and India favoring operators willing to keep training data and inference workloads within national borders and to partner with local utilities on power access.