NVIDIA CEO Jensen Huang on DeepSeek: Why Investors Got It Wrong

NVIDIA (NASDAQ: NVDA) CEO Jensen Huang addressed investor concerns about DeepSeek and its open-source reasoning model, emphasizing that the market misinterpreted its impact on NVIDIA and the AI industry.

The DeepSeek Sell-Off and Market Reaction

• NVIDIA's stock dropped 17% on January 27 after DeepSeek made headlines. However, shares have since recovered most of those losses.
• Investors assumed DeepSeek's advancements would reduce demand for NVIDIA's AI hardware, driving concerns over its long-term growth prospects.
• Huang argues that the market misunderstood the AI training process and overestimated the negative impact.

Huang's Perspective: Why DeepSeek Isn't a Threat

1. The Misunderstood Paradigm

Huang explains that investors wrongly assumed AI training was a two-step process—pre-training followed by inference. However, he clarifies that post-training is an equally critical step.

• Pre-training establishes the foundation model.
• Post-training is where the real learning happens—this phase is computationally intensive, requiring substantial processing power.
• Investors overlooked the complexity of post-training, which continues to drive demand for NVIDIA's AI chips.

2. The Importance of AI Reasoning

• Huang explains that more reasoning before answering a question leads to better AI performance.
• Reasoning is compute-intensive, meaning AI models will continue to require massive processing power, benefiting NVIDIA's GPUs.
• He dismissed concerns that DeepSeek signals the end of AI innovation, calling it “exactly the opposite.”

What's Next for NVIDIA?

Despite short-term volatility, NVIDIA remains at the forefront of AI hardware development, with strong demand for its GPUs in AI training, post-training, and reasoning.

For a detailed financial breakdown of NVIDIA, including earnings reports, financial statements, and valuation metrics, check out the Full Financials API from Financial Modeling Prep.