According to Nvidia’s CEO, the impact of Chinese AI company DeepSeek was overestimated, and chips will continue to play an important role within artificial intelligence.
Jensen Huang, CEO of Nvidia, expressed his views on Chinese AI company DeepSeek during an interview with Alex Bouzari, CEO of DataDirect Networks. Although this new model implied a dip in shares of Nvidia’s chips, Huang is positive about DeepSeek’s R1 model. “I think the market reacted to R1 with: ‘Oh my god. AI is ready.’ ‘ It came out of the blue. We didn’t have to do any more computation.’ But no, it’s completely the opposite,” Huang said during the interview.
Adoption AI
According to Jensen, the arrival of DeepSeek’s R1 model would just contribute positively to the adoption and wider acceptance of AI. “It makes everyone notice that there are opportunities for models to be much more efficient than we thought was possible,” Huang said.
read also
AI computing power grows exponentially according to Nvidia CEO
In addition, he says that despite the progress DeepSeek has made in pre-training AI models, post-training remains an important and labor-intensive process. “Reasoning is a pretty computationally intensive part of it,” he says. By this, he wants to emphasize that post-training is important and that chips and computers will be needed.
Shares of Nvidia
About a month ago, Chinese AI company DeepSeek surfaced with the R1 model, which would perform better for a lower price compared to other established models. This momentarily rocked the market, as well as Nvidia’s shares. The company’s share price fell 16.9 percent one market day after DeepSeek’s launch.
read also
Inside DeepSeek-AI stirs minds: how does it work, and what’s stolen?
The Chinese AI company thereby swept some $600 billion off the table for Nvidia. Meanwhile, the chip company’s shares are said to have almost fully recovered.