Europe and South Korea Race to Build Sovereign AI Systems

Europe and South Korea are both accelerating their efforts to build sovereign artificial intelligence capabilities as concerns mount over dependence on American and Chinese technology, with industry leaders warning that nations without homegrown AI systems risk being left behind in the rapidly approaching era of artificial general intelligence. Guillaume Lample, chief scientist and co-founder of Mistral AI, Europe's leading AI startup, issued a stark warning that Europe must urgently develop its own superintelligence, cautioning that commercial or geopolitical rivals could withhold life-saving technologies from the region if it lacks independent systems. Speaking ahead of a company event in Paris, Lample said the arrival of AGI is imminent and emphasized that without access to these models in Europe, the consequences could be severe. Meanwhile, South Korea's Deputy Prime Minister and ICT Minister Bae Kyung-hoon announced that Korea should now attempt to build frontier AI models comparable to those of the US and China, moving beyond the country's current focus on industrial applications in semiconductors and manufacturing. Bae acknowledged that Korea's entire government AI budget equals only what a single US technology company invests, but suggested that more aggressive GPU and infrastructure investment could enable the challenge. Mistral AI, founded by three French researchers who previously worked at Google and Meta, positions itself as a practical, business-focused player offering AI models hosted entirely on European data centers. However, CEO Arthur Mensch identified funding as the company's biggest hurdle, noting that unlike US giants that can spend tens of billions ahead of demand, Mistral must borrow money based on signed contracts. In a separate initiative, the Pincus Fund and partners are developing JEDAI, a nonprofit AI infrastructure designed specifically for Jewish education, backed by a one million dollar Israeli Innovation Authority grant. The platform, launching global pilots in January 2027 across North America, Latin America, Europe, and South Africa, aims to ensure that AI tools reflect values around identity and community rather than generic commercial efficiency.