The Real Threat of Chinese AI
In February 2025, the emergence of DeepSeek, a Chinese AI startup, began to reshape the global AI landscape. Developed in Beijing, DeepSeek’s rapid rise underscores a pivotal moment in the global AI race. The United States must recognize the urgency and adapt its strategies to maintain its leadership position.
The Open-Source Advantage
DeepSeek’s success results, in part, from its embrace of open-source software. This model allows anyone to download, modify, and build upon its foundations. Unlike most comparable American AI models, which are proprietary and restrict use, DeepSeek’s open-source nature has driven its rapid adoption. This open approach has a history of bolstering innovation and technical progress. The emergence of open-source large language models (LLMs), such as Meta’s Llama and DeepSeek, parallels the co-existence of the Linux operating system alongside proprietary systems such as Microsoft’s Windows. This allows freedom in access and modification of code, similar to AI’s ‘Linux moment’.
AI With Chinese Characteristics
The impact of DeepSeek’s rapid growth extends beyond its technical achievements. Its use could give China the power to embed its values into the very fabric of generative AI. Beijing requires Chinese-made LLMs to align with the “core values of socialism” and restrict “problematic” content. If you ask DeepSeek-V3 about the 1989 Tiananmen Square massacre, for example, it replies, “I am sorry, I cannot answer that question.” This censorship poses a significant threat to the free flow of information and could allow the spread of Chinese propaganda. Moreover, DeepSeek’s popular app sends U.S. user data directly to China.
A New American Strategy
To counter the potential threat of Chinese AI, the U.S. must take several steps. These strategies include developing and distributing U.S.-built open-source LLMs via universities, companies, and national labs. Also, Washington should incentivize investment in open-source AI systems that are compatible with Western chipsets. The government should also consider partnerships with initiatives like Stargate, a collaboration investing in new AI infrastructure in the United States.
To ensure American dominance, the United States must lead the next wave of AI innovation. Furthermore, Washington should fund next-generation model development and support efforts to design and produce optimized hardware. The government should also encourage firms making big bets on new software and hardware approaches that will underpin tomorrow’s leading-edge AI systems.
Lead or Lose
The United States stands at a critical juncture. It must quickly implement a comprehensive strategy to build and maintain its leadership in open-source AI. The alternative of not prioritizing open-source AI would endanger U.S. influence over AI systems. The U.S. should embrace its history of leading open models while ensuring a competitive ecosystem and continuing to invest in critical resources, whether they are chips or human talent. Second place is not an option.
