Chinese AI startup DeepSeek is preparing to launch V4, its first multimodal model since R1 wiped nea...
The AMW Read
Updates the DeepSeek case study (01.§4) with a multimodal roadmap that explicitly signals US-CN decoupling (cross.§E) through the decoupling of model development from Nvidia/AMD silicon (cross.§H).
Chinese AI startup DeepSeek is preparing to launch V4, its first multimodal model since R1 wiped nearly $600 billion off Nvidia's market value in January 2025. The new model processes text, images, and video while running on Chinese chips from Huawei and Cambricon, deliberately excluding Nvidia and AMD from early access. This strategic move demonstrates China's accelerating tech independence despite US export controls and could reshape global AI infrastructure assumptions. The timing coincides with China's parliamentary sessions on March 4, signaling the geopolitical significance of this launch.
