The term “deep-six,” originating from nautical terminology in the 1920s, refers to the act of throwing items overboard into a six-foot watery grave, symbolizing an end or disposal of something. Today, this term is being invoked in discussions about the potential disruption caused by China’s Deepseek, which claims to have developed an advanced AI large language model that could rival the best offerings from the United States for a fraction of the cost and electricity. This has raised concerns that Deepseek’s innovations could lead to the decline of several major tech and utility companies in the AI sector.
As the news surrounding Deepseek continues to unfold, the immediate market reactions have been significant. Following Deepseek’s announcement, some leading AI companies experienced stock price declines of 15% to 25% in a single day. Notably, Nvidia faced its largest one-day loss in history, with an 18% drop and a 24% decline from its all-time highs. However, such fluctuations are not unprecedented. For instance, Nvidia previously experienced a 35% drop last summer, which did not deter long-term investor confidence nor disrupt its overall upward trajectory.
The critical question now is whether Deepseek’s entry into the market will have lasting implications for the demand for high-end chips and AI processing power. Should Nvidia’s stock fall below the 90s range, it could signal significant technical damage and a potential shift in market dynamics. This would imply that the availability of cheaper AI models from Deepseek might reduce the demand for high-performance hardware, which has been the backbone of AI development. However, it’s likley Deepseek and Alibaba’s cheaper AI models are built upon US platforms with pilfered chips. Deepseek is dated, lacks mulit-modal and not useful for major applications in robotics and automomous. Yet, the prospect of cheaper AI training and inference platforms have taken the emotional premium off of Nvidia’s prior growth trajectory.
Interestingly, while the initial response from the market has been volatile, many investors seem to view the potential decrease in costs associated with AI training and inference models as a positive development, particularly for software stocks. The prospect of lower operational costs could foster innovation and accessibility in AI applications, potentially benefiting the software sector even if demand for high-end hardware declines. The tech sector quickly recovered with some segemnts already back to record highs.
In conclusion, while the emergence of Deepseek poses challenges for established players in the AI space, it also opens up new opportunities. The shift in investor focus from hardware to software indicates a broader trend where the value of AI applications may take precedence over the underlying infrastructure. As the situation develops, it will be crucial to monitor how the competitive landscape evolves and how companies adapt to these changes in the AI sector.
Far from deep-sixing the stock market, it appears that the Mag 7 leaders and the broader market have taken the news of cheap Large Language training models in stride. The SP is already back to testing record highs while moribund sectors such as small, mid cap and healthcare are leading in 2025, joining our other favorite sectors: financials, cybersecurity and bitcoin.
