Jensen Huang heaped reward on the Chinese language AI mannequin that briefly upended the tech world, calling DeepSeek’s R1 “a terrific contribution to the business and to the world” on Monday.
Shares of tech and semiconductor firms, together with Nvidia, tumbled in January following the meteoric rise of DeepSeek R1, the Chinese language AI mannequin that traders seen as being globally aggressive and cost-effective.
However Huang has good issues to say about DeepSeek, which he mentioned on Monday was “genuinely a present to the world’s AI business.”
“The quantity of pc science breakthroughs is admittedly fairly vital and has actually opened up lots of nice analysis for researchers in the US and around the globe,” Huang mentioned on the opening keynote of the Computex Taipei tech convention in Taiwan.
In January, open-source chatbot DeepSeek R1 took the world by storm, elevating questions on Silicon Valley’s huge spending spree on the know-how.
“In all places I am going, DeepSeek R1 has made an actual impression in how folks take into consideration AI and the way to consider inference and the way to consider reasoning AIs,” Huang mentioned.
US AI-related shares tanked throughout the board within the wake of DeepSeek’s rise. Nvidia’s inventory misplaced as a lot as $600 billion in market capitalization, hitting 20% of Huang’s private internet value at one level. The inventory has recovered most of those losses and is up almost 43% within the final 12 months.
Huang mentioned in February that traders bought it improper as a result of the business will nonetheless want computing energy for post-training.
On the time, Huang mentioned that post-training is the “most necessary a part of intelligence” and “the place you be taught to resolve issues.”
The tech titan additionally appeared upbeat about DeepSeek, saying the open-sourced mannequin created “power around the globe.”