China's New Breakthrough a Disaster for America, Nvidia Boss Speaks Out

Jakarta, CNBC Indonesia - Nvidia CEO Jensen Huang has spoken out regarding the collaboration between DeepSeek and Huawei, which has introduced the latest AI model. According to him, China’s push for self-sufficiency and US export controls will impact the chip giant.

DeepSeek is known to have launched a preview version of its latest V4 model. This launch also signifies collaboration with Huawei, as the company previously used Nvidia.

Huawei chips were used in some of the V4 training processes. “The day DeepSeek was first launched on Huawei is a terrible outcome for our country,” Huang said, quoted from Reuters, on Monday (27/4/2026).

The model resulting from the DeepSeek and Huawei collaboration is named V4. The pro version of the model is claimed to surpass other open-source models, only trailing Google’s closed-source Gemini Pro-3.1.

“This is a big deal for China’s AI industry,” said He Hui, director of semiconductor research at consultancy firm Omdia.

Meanwhile, machine learning engineer at Hugging Face, Lewis Tunstall, explained that V4 has become the fastest model at the top position on that developer forum for sharing and running machine learning.

AI engineer who has tested V4, Daniel Dewhurst, said the model preview is very promising. However, he is still awaiting tests conducted by other developers around the world.

“My initial view is that the DeepSeek V4 preview looks significant, but I would be cautious about accepting benchmark headlines at face value until there is more independent evaluation and real-world testing by developers,” he explained.

DeepSeek explained that V4 is suitable for AI agent work, so it can handle more complex tasks than chatbots with greater computation.

However, DeepSeek has not provided a schedule for when V4 will be launched.