DeepSeek V4 AI Model Set for Launch, Posing New Challenge to US Tech Giants
DeepSeek, the Chinese artificial intelligence startup that previously sent shockwaves through global markets by wiping nearly $600 billion off Nvidia's market value in a single day with its R1 model, is reportedly on the brink of unveiling its next major AI innovation. According to a report by The Financial Times, the company is preparing to launch the V4 model as early as next week, marking a pivotal moment for China's technology sector as it strives to demonstrate its ability to compete with American industry leaders despite stringent US restrictions on high-performance computing chips.
What is DeepSeek V4 and Its Potential Impact on US Companies
Unlike its predecessors, the V4 model is anticipated to be "multimodal," meaning it will not only process text but also comprehend and generate images and videos, similar to capabilities seen in Google's Gemini 3.0. This development represents DeepSeek's first significant move since January 2025, when its R1 model astonished the world by claiming to achieve AI intelligence comparable to top Silicon Valley models while using only a fraction of the power and cost. That assertion triggered investor panic, leading to fears that Nvidia's expensive chips might no longer be essential for building powerful AI systems, resulting in a dramatic stock decline for Nvidia.
DeepSeek has reportedly optimized the V4 model to operate on domestically produced hardware from Chinese companies such as Huawei and Cambricon, in contrast to American giants that rely heavily on Nvidia chips. By leveraging local chip technology, DeepSeek is effectively circumventing US export controls that block China from purchasing the latest processors from Nvidia and AMD. This strategic move is expected to hasten China's transition away from Western technology for AI "inference"—the critical process where an AI model generates responses for users.
Controversy and Competition Surrounding DeepSeek
DeepSeek's rapid ascent has not been without controversy. Recently, the US AI firm Anthropic accused DeepSeek of engaging in "distillation attacks," a practice where a company uses outputs from a more advanced AI, like Anthropic's Claude, to train its own smaller model. Additionally, while DeepSeek was previously lauded for its openness in sharing engineering techniques, it is now expected to adopt a more guarded approach with the V4 launch, releasing only a brief technical note alongside the model.
This upcoming release underscores the intensifying global AI race, with DeepSeek positioning itself as a formidable contender against established US players. The success of V4 could further destabilize market dynamics, particularly for chip manufacturers like Nvidia, and reinforce China's ambitions in the AI domain.
