Quarterly Outlook
Macro outlook: Trump 2.0: Can the US have its cake and eat it, too?
John J. Hardy
Chief Macro Strategist
Chief Investment Strategist
DeepSeek, a Chinese artificial intelligence (AI) startup, has recently emerged as a significant player in the AI landscape, drawing considerable attention from global investors and industry leaders. Founded in 2023, DeepSeek has rapidly developed advanced AI models that rival those of established U.S. tech giants.
DeepSeek unveiled a free, open-source large-language model in late December that it says took only two months and less than $6 million to build, using reduced-capability chips from Nvidia called H800s. DeepSeek’s app is now one of the most downloaded app on Apple’s App Store, and Bloomberg searches for “DeepSeek” have surged, reflecting heightened global interest and investor focus.
This has raised a few pertinent questions, such as:
DeepSeek’s AI models, like the DeepSeek-R1, have demonstrated superior performance compared to prominent U.S. models such as OpenAI’s GPT-4 and Meta’s LLaMA 3.1. These achievements span across various benchmarks, including complex problem-solving, mathematics, and coding, positioning DeepSeek as a key player in the global AI arena.
Even if DeepSeek does not maintain its current level of popularity, this development serves as a reminder that competition in the global AI arena is intensifying, and Nvidia may not be in the pole position forever.
By developing cutting-edge AI models with less advanced and more cost-efficient hardware, DeepSeek challenges the heavy investments U.S. tech companies are pouring into high-cost AI infrastructure.
Reports suggest that DeepSeek-R1’s API costs just $0.55 per million input tokens and $2.19 per million output tokens, compared to OpenAI’s API, which costs $15 and $60, respectively. This raises critical questions about whether the traditional capital-intensive approaches are sustainable in the long run.
Taking a different path from many U.S. companies that tightly guard their AI models, DeepSeek has adopted an open-source strategy, openly sharing its code and training methodologies.
This approach could drive faster innovation and broader adoption globally, potentially diminishing the competitive advantage of U.S. firms reliant on proprietary systems such as OpenAI’s GPT-4 or Google’s Bard. Yann LeCun, Meta's chief AI scientist, remarked that DeepSeek's success underscores the potential of open-source models to surpass proprietary ones.
U.S. tech companies are trading at premium valuations, with major AI players like Nvidia, Microsoft, and Alphabet commanding forward P/E multiples far above historical averages. With these stocks priced for perfection, even minor disruptions, such as DeepSeek proving advanced AI can be built without top-tier chips, could weigh heavily on share prices. For Nvidia, in particular, its role as a key supplier of AI chips makes it vulnerable if demand for its high-end products wanes.
In contrast, Chinese tech companies, including new entrants like DeepSeek, are trading at significant discounts due to geopolitical concerns and weaker global demand. DeepSeek's rise could spark renewed investor interest in undervalued Chinese AI companies, providing an alternative growth story.
DeepSeek’s success raises another pressing question: how did they do it despite U.S. export controls? The Biden administration’s semiconductor restrictions were designed to limit China’s access to high-powered chips like Nvidia’s H100s. Yet DeepSeek has managed to thrive, suggesting either a workaround or that these controls aren’t as restrictive as intended.
This could have major implications for both U.S. policy and investor sentiment toward Chinese tech. If companies like DeepSeek can innovate despite sanctions, the narrative around China's technological dependence on the U.S. may shift dramatically.