NVIDIA Corp (NVDA)

[rolvram]

NVIDIA’s CEO Jensen Huang Addresses The “DeepSeek Fiasco” For The First Time; Says Investors “Overreacted” & “Got It All Wrong”
Muhammad Zuhair

Feb 21, 2025 at 08:31am EST


NVIDIA's CEO Jensen Huang has publicly addressed China's DeepSeek for the first time, claiming that the investors overreacted and got the "paradigm wrong."

NVIDIA's CEO Believes DeepSeek's AI Achievement Is Massive For The Markets, Appreciating Their Open-Source Models
The release of DeepSeek's R1 AI models was a massive milestone for the AI markets, but for Team Green, it was an ugly day. The firm saw a whopping $600 billion decline in market value, with Jensen losing over 20% of his net worth, clearly showing investors weren't happy with DeepSeek's achievement. However, in the latest interview with DDN, NVIDIA's CEO Jensen Huang has expressed excitement towards DeepSeek's milestone and, at the same time, believes that investors' perception of AI markets went wrong.

From an investor perspective, there was a mental model that the world was pre-training and then inference. And inference was: you ask an AI a question, and you instantly got an answer. I don't know whose fault it is, but obviously that paradigm is wrong.

It is so incredibly exciting. The energy around the world as a result of R1 becoming open-sourced, incredible.

For those who still aren't aware of why the stock sell-off got triggered, the news around DeepSeek's R1 being trained for around $5 million raised the perception that the demand for AI computing power is artificial in the markets. However, DeepSeek's "low-training" costs were only a FUD, and it was reported that DeepSeek employs well over $1 billion in AI hardware, showing that the firm, too, needs massive computing power.

NVIDIA Blackwell GB200 Powers The Latest A4 VMs At Microsoft Azure, 2.25x Higher Compute & HBM Capacity 1
However, one area where DeepSeek managed to tap into is having robust "open-sourced" AI models, which means that developers can join in to enhance the product further, and it allows organizations and individuals to fine-tune the AI model however they like, allowing it to run on localized AI environments and tapping into hardware resources with the best efficiency. Prior to DeepSeek, the perception was general against open-sourcing models, mainly due to the fact that OpenAI drove the hype.

All eyes are on NVIDIA's upcoming earnings call, which is slated for February 26. The call will likely give us insight into how big of a hit the firm has seen on profitability rates following the DeepSeek fiasco and recent Blackwell AI product issues. However, it is safe to say that with competition from DeepSeek, it is certain that demand for computing power is all around NVIDIA.