China’s DeepSeek introduced its R1 model, claiming training costs roughly 70% below comparable Western AI systems. The multimodal model supports text, image, and speech generation, optimized for enterprise deployment across cloud and edge devices. CEO Liu Yun said cost efficiency and rapid fine-tuning make R1 attractive for developers in emerging markets. Industry experts noted that DeepSeek’s scaling approach challenges U.S. AI incumbents on price performance, potentially broadening AI adoption among mid-tier firms globally as compute costs remain a bottleneck.