Chinese AI company DeepSeek has introduced a preview version of its new AI model called V4, making China less dependent on foreign tech industries. The model was designed to shift away from the heavy use of Nvidia hardware.
According to DeepSeek, the V4 Pro model performs better than most open-source AI models and is only behind Google’s Gemini-Pro-3.1. This step is considered an important step for China’s AI industry because Huawei’s Ascend chips are considered the strongest alternative to Nvidia processors.
Previously, DeepSeek mainly relied on Nvidia chips for training its AI systems. Now, Huawei confirmed that its Ascend chips were used during parts of the V4 training process. This partnership shows that advanced Chinese AI models can now run successfully on Chinese-made hardware.
The launch also highlights the ongoing technology competition between the United States and China. NVIDIA CEO Jensen Huang recently warned that US export restrictions could push Chinese developers toward domestic technology instead of American products.
The new V4 model quickly became popular as it uses less computing power than many competing systems.
However, unlike some leading AI models, V4 currently cannot process images or videos. DeepSeek has faced backlash from US officials and competitors, claiming the company may have benefited from American technology by using Nvidia chips in the past but has denied intentionally using AI-generated data.
At the same time, Chinese chip companies such as SMIC and Huahong Semiconductor saw their shares rise after investors expected stronger demand for local chips.
Consequently, DeepSeek is helping China move towards the greater tech independence and making the country less dependent on foreign countries. It is an important step in the global AI competition. Reports also suggest the company is seeking new investments that could value it at more than $20 billion.
Also read: OpenAI officially launches ChatGPT-5.5
