In a significant development for the AI industry, NVIDIA CEO Jensen Huang has finally addressed the recent market turbulence caused by Chinese AI company DeepSeek's breakthrough models. His response comes after NVIDIA's stock experienced notable volatility following DeepSeek's introduction of cost-effective AI models.
DeepSeek Recent Releases:
- DeepSeek-V3
- DeepSeek-R1 Notable features: Low-cost training, High cost-performance ratio
Market Impact and Initial Response
In a conversation with DDN Storage CEO Alex Bouzari on February 20th, Huang directly addressed the DeepSeek effect that caused NVIDIA's stock to plummet 17% in a single trading day. He emphasized that investors had fundamentally misinterpreted the implications of DeepSeek's emergence in the AI landscape, particularly regarding computational requirements.
Key Events Timeline:
- February 20, 2025: Jensen Huang's first public response to DeepSeek
- NVIDIA stock impact: 17% single-day decline
- February 26, 2025: Scheduled NVIDIA Q4 earnings report
Technical Reality vs Market Perception
Huang strongly contested the market's assumption that DeepSeek-R1's arrival would eliminate the need for intensive computation. He explained that the opposite is true - inference remains computationally intensive, and the quest for higher-quality AI responses actually demands more, not less, computing power. The NVIDIA CEO highlighted that post-training phases are crucial for AI development and continue to require substantial computational resources.
DeepSeek's Impact on AI Industry
Despite the market's initial negative reaction, Huang praised DeepSeek's innovations, particularly the DeepSeek-R1 inference model, describing it as impressive. He acknowledged how the R1 model has energized the global AI community, while emphasizing that its sophisticated reasoning capabilities actually drive increased demand for computing power.
Industry Perspectives
The discussion has extended beyond NVIDIA, with AMD CEO Lisa Su offering her perspective during recent earnings calls. Su views DeepSeek's innovations as catalysts for AI democratization, noting that while new training and inference methods may require less infrastructure, they ultimately contribute to expanding AI applications and deployment scenarios.
Future Implications
As NVIDIA prepares to release its fourth-quarter earnings report on February 26th, the industry continues to watch how this technological evolution will shape the future of AI computing. Huang's response suggests that rather than threatening existing hardware providers, innovations in AI efficiency may actually expand the overall market for computational resources.