Governments Turn to Open-Weight AI to Bolster Tech Sovereignty

Open-weight models can be downloaded, modified and run on infrastructure controlled by their users, offering governments and companies an alternative to proprietary APIs whose prices, access rules and data locations are set by outside vendors. That flexibility has become strategically important as artificial intelligence moves into critical services. China, the European Union and Canada are increasingly treating models, computing capacity and data governance as elements of technological sovereignty, not merely procurement choices.

Mozilla released its inaugural State of Open Source AI report on July 14, 2026, drawing on a survey of more than 950 developers. It said the performance gap between leading open and proprietary models had narrowed to 3%, while inference costs fell as much as 50-fold over three years. Adoption reached 89% in Greater China and East Asia. France has committed 109 billion euros to AI infrastructure, while Canada allocated C$890 million for a sovereign public supercomputer; 47 countries restrict foreign processing of critical workloads.