Why is the U.S. worried about DeepSeek's rise?

The rise of Chinese AI company DeepSeek has sparked heated global discussions in recent days. Surging to the top of the free app download charts in the U.S. Apple App Store – overtaking the once-dominant ChatGPT – it has also triggered a sharp decline in the stock prices of several American tech firms.

On Monday, U.S. President Donald Trump called DeepSeek's emergence a "wake-up call" for American industries. Meanwhile, U.S. congressional offices were informed that the use of DeepSeek was "unauthorized for official House use," Axios reported Thursday.

Why is DeepSeek's rise causing such concern – even anxiety – in the U.S.? A commentary by Yuyuan Tantian, a social media account affiliated with China Media Group, offers three key reasons, citing insights from the China Academy of Industrial Internet.

1. Cost-effective disruption to the U.S. AI model monopoly

DeepSeek is redefining efficiency in AI development, delivering top-tier performance at a fraction of the cost. While OpenAI reportedly spent $78 million training GPT-4, DeepSeek achieved similar results with under $6 million. By pioneering a more efficient training approach, it has lowered AI's entry barriers, making large-scale pre-training accessible beyond tech giants.

Additionally, DeepSeek's newly launched DeepSeek-R1 is priced at just $2.2 per million tokens – compared to OpenAI's o1 model at $60 per million tokens. This breakthrough makes AI more affordable, unlocking vast potential for research institutions, enterprises and knowledge-intensive industries.

With disruptive advancements in both training and inference, DeepSeek is reshaping the AI landscape, setting new standards for efficiency and accessibility while challenging traditional models.

2. Innovative approach sparks panic among U.S. tech professionals

DeepSeek has drastically reduced development costs through a novel training strategy. Unlike OpenAI, which relies on massive computational power, DeepSeek maximizes efficiency by using advanced algorithms to filter, summarize and selectively process training data. This method significantly cuts costs while improving performance.

The impact is already being felt. Despite Meta pouring vast resources into training its AI model Llama, it has failed to outperform DeepSeek, which operates at a fraction of the cost. This has prompted Meta executives to question their spending and efficiency, fueling anxiety among U.S. tech professionals who fear their jobs and expertise may be at risk.

3. China's AI models are gaining momentum

According to the China Academy of Industrial Internet, from Q4 2023 to Q1 2025, the gap in AI model capabilities between China and leading global firms has narrowed by nearly 75 percent. This suggests that DeepSeek's rise is not an anomaly but part of China's broader strategic advancements in AI.

The report also highlights global AI investment trends, with China securing the second-largest AI investment at $5.5 billion – compared to the U.S.'s $64.1 billion. While this indicates that the U.S. still leads in funding, it also underscores China's significant room for future growth in the AI sector.