Why the Deepseek model needs “100 times more computer”

CEO Jensen Huang discussed the impact From the Chinese startup of the new Deepseek artificial intelligence model in an interview on Wednesday with CNBC Jim Cramer has NvidiaAnnual GTC conference. Huang said the revolutionary model would really require No more calculation that many in the industry thought.

According to Huang, Deepseek’s R1 model is “fantastic” because it is “the first open source model”. He explained that the model breaks down the problems step by step, is able to find different answers and can check if his answer is correct.

“This reasoning consumes 100 times more calculation than a non-season AI,” he said. “It was exactly the opposite, it was exactly the opposite conclusion that everyone had.”

At the end of January, the Deepseek model sparked a massive sale in AI shares, because investors feared that the model could work as well as the best competitors using less energy and money. Nvidia fell 17% in a session to lose nearly $ 600 billion, the the biggest ever A decrease of one day for an American company.

Huang also talked about some of its companies revealed Until now at its conference, including new AI infrastructure for robotics and the company, highlighting partnerships with companies like Apron,, Hpe,, Emphasis,, Serve And Cowsterrike. He thinks widely on the Boom of AI, noting the way in which media threw has gone from purely generative AI to reasoning models. He also predicted that the IT capital expenses in the world seem to be on the right track to reach a dollars Billion by the end of the decade, and that the majority of this money will be used for AI.

“So our percentage opportunity of a dollars Billion by the end of this decade is quite large,” said Huang. “We have a lot of infrastructure to build.”

Register So that the CNBC Investing Club can follow each Jim Cramer movement on the market.

Non-liability clause The CNBC Investing Club Charitable Trust has actions from Nvidia and Crowstrike.