NVIDIA, the largest supplier of chips used to train artificial intelligence software, has said the new model launched by Chinese startup Deepseek is an “excellent AI advancement” that complies with US technology export controls.
“Deepseek’s work illustrates how new models can be created” using the test Time Scaling technique, the company said in an email statement, “taking advantage of widely available models and computing that is fully in accordance with export controls. ”
Nvidia’s statement indicates that the company believes that the Chinese company has not violated US restrictions that limit US access to advanced chips in creating its technology. Most NVIDIA’s best products cannot be sold to Chinese companies without Washington licenses – permissions that are usually not granted.
Read more:
The statement also seems to disregard some suspicions of analysts and experts that Chinese startup could not have made the advancement it claims to have reached.
After more than quintupplicated its value in the last two years, Nvidia’s shares plummeted on Monday (27) in one of the biggest sales in history, due to the concern that DeepSeek’s supply reduces the need for dear hardware type that US chips company offers. However, Nvidia’s statement points out that inference, the work of performing AI models, still requires many of its products.
Continues after advertising
“Inference requires a significant number of NVIDIA GPUs and high performance networks,” said the company.
© 2025 Bloomberg L.P.
