![]() |
| Credit to Nvidia |
The AI world was shaken when DeepSeek unveiled its R1 open-source reasoning model, sending Nvidia’s stock into a freefall. But according to Nvidia CEO Jensen Huang, the market completely misread the situation.
Speaking in a pre-recorded interview with Alex Bouzari, CEO of DataDirect Networks, Huang called DeepSeek’s innovation “incredibly exciting” and rejected the notion that it threatens Nvidia’s dominance.
“I think the market reacted as if AI was over,” Huang said. “Like, ‘Oh my gosh, we don’t need computing anymore!’ But it’s exactly the opposite.”
Instead of spelling doom for the AI hardware market, Huang believes DeepSeek’s advancements will drive even greater demand for AI processing power.
“It’s making people realize that AI models can be far more efficient than we ever thought possible,” Huang explained. “That means AI adoption is going to expand and accelerate.”
He also emphasized that while DeepSeek’s pre-training techniques are impressive, the post-training process—where AI models are refined and fine-tuned—remains highly compute-intensive.
“Reasoning is still a major computational challenge,” he added, hinting that Nvidia’s chips will continue to play a crucial role in AI’s evolution.
Market Overreaction?
Huang’s comments come nearly a month after DeepSeek’s R1 announcement sent shockwaves through the market. Nvidia’s stock plunged 16.9% in a single day, erasing a staggering $600 billion from its market cap within three days.
However, Nvidia has since bounced back. On Friday, its stock opened at $140 per share, nearly recovering from the losses. With the company’s Q4 earnings report set for February 26, investors will be eager to hear more about Nvidia’s stance on DeepSeek and the future of AI computing.
Meanwhile, DeepSeek isn’t slowing down. The company just announced plans to open-source five additional code repositories next week as part of its “Open Source Week” event.
One thing is clear: AI is far from “finished.” If anything, the race is just heating up.



:max_bytes(150000):strip_icc():format(webp)/GettyImages-1258889149-1f50bb87f9d54dca87813923f12ac94b.jpg)