Nvidia Stock Drops Amid AI Market Concerns
Overview
Nvidia's (NVDA) stock experienced a significant decline of over 12% in early trading on Monday. This sell-off was triggered by the release of a new AI model by Chinese startup DeepSeek, which raised concerns about AI investment and the potential for more cost-efficient artificial intelligence agents.
DeepSeek's AI Model
On January 20, DeepSeek released a new AI model that has been perceived as a potential threat to OpenAI. Venture capitalist Marc Andreessen praised the model's advancements, describing it as "one of the most amazing and impressive breakthroughs" he had witnessed.
Training Costs and Chip Demand
DeepSeek's model reportedly cost $5.6 million to train, significantly less than OpenAI's GPT model, which cost over $100 million. This cost difference raises fears that AI models may start requiring fewer chips and energy, potentially impacting the demand for Nvidia's high-end chips.
Chip Stock Sell-Off
The news of DeepSeek's model sent shockwaves through the chip industry, resulting in a broad sell-off across chip stocks. Broadcom (AVGO) declined over 14%, Micron (MU) lost almost 8%, and Advanced Micro Devices (AMD) fell nearly 5%.
Analyst Perspectives
Analysts have offered differing opinions on the implications of DeepSeek's model. While Raymond James analyst Srini Pajjuri suggests that training costs could decrease, Bernstein analyst Stacy Rasgon cautions against overreacting, noting that DeepSeek's announced cost "does not include all the other costs associated with prior research."
US Export Restrictions
Additionally, tightened US export restrictions implemented during President Biden's final days in office could hinder DeepSeek's ability to continue training new models by limiting China's access to Nvidia's AI chips and chipmaking machines.
US AI Investment
In contrast to the concerns surrounding DeepSeek's model, the US is ramping up its investment in AI infrastructure. President Trump recently announced the Stargate AI project, a $100 billion initiative to build data centers and other AI-related infrastructure.