Palantir CTO Warns Chinese Open-Source AI Models Leverage US IP, Calls for Enhanced Protection Measures

Sham Saunk, Palantir's Chief Technology Officer, warned that Chinese open-source AI models are utilizing US intellectual property (IP) and urged stronger countermeasures. Speaking at a defense innovation conference in Pennsylvania, Saunk claimed that many Chinese AI models are developed through 'distillation'—a process where smaller models are trained using outputs from larger ones. While distillation is widely used in the industry for its cost-efficiency, Saunk alleged that Chinese firms are improperly leveraging US research. He emphasized the need for US AI firms to strengthen IP protections, though debates persist over whether distillation constitutes IP theft. Meanwhile, Saunk acknowledged France's decision to prioritize local companies over Palantir, framing it as part of Europe's push for tech sovereignty. The evolving US-China AI rivalry now centers on data governance and IP protection, raising questions about potential regulatory responses.