Summary of Key Points
This article corrects the misconception that "China started AI late" by comparing the development paths of artificial intelligence (AI) in China and the United States. It highlights significant differences between the two countries in terms of technological strategy, ecosystem building, talent mobility, and goal orientation: The U.S. focuses on "raising the technical ceiling" (cutting-edge models, high-end chips, long-term investment), while China excels at "lowering the application barriers" (open-source technology, low costs, and industrial implementation). The U.S. aims for artificial general intelligence (AGI), whereas China is more concerned with transforming AI into practical productivity. Additionally, although China is a major producer of AI talent, the U.S. remains a hub for talent aggregation. The competition between the two countries is not just about technological disparities but also about strategic choices in their respective ecosystems.
1. Chinese AI Didn't Start “Decades Later” – You Just Didn’t Notice
Many people think that China’s AI efforts began only after the success of ChatGPT, but this is incorrect. As early as the Dartmouth Conference in 1956, China was already conducting research in areas such as automatic control, speech recognition, and machine translation; however, these developments were confined to laboratories and not referred to as AI by the general public. For example, Zhang Yiming’s Toutiao (a news recommendation platform) in 2012 used algorithms, which is a form of AI application, but at that time, it was simply called "algorithm." It wasn’t until the emergence of ChatGPT in 2023 that AI moved from the background to the forefront and became a topic of widespread discussion.
At the national level, China made early plans: In 2017, the State Council issued the "New Generation Artificial Intelligence Development Plan," with the goal of becoming a leading global AI innovation center by 2030. Therefore, China’s AI development is not something that started suddenly but has been a decades-long process, just recently coming to the attention of the public.
2. The U.S. Raising the “Ceiling,” China Lowering the “Barriers”
The core difference between Chinese and American AI can be summarized in one sentence: The U.S. continuously pushes the technical limits of AI higher, while China works to lower the barriers to its use.
- The U.S.’s “Ceiling”: Companies like NVIDIA control the AI chip and software ecosystems, Microsoft/Google dominate cloud computing, and OpenAI develops cutting-edge models. They are willing to invest billions of dollars in research, such as in artificial general intelligence (AGI). They are not afraid of failure and allow experts to make mistakes over the long term; a few successes can lead the way.
- China’s “Barriers”: Chinese companies focus on making AI more affordable and user-friendly. For instance, DeepSeek’s open-source models are freely downloadable, and AI has been integrated into smartphones, cars, and factories, making it accessible to a wider audience. Chinese consumers are accustomed to free services, so companies must monetize through scale and specific use cases, which drives the popularization of AI—even if they don’t lead the world, as long as it’s affordable and easy to implement.
The ceiling determines the height of technology, while the barriers determine its reach. In history, many technological revolutions (such as cars and smartphones) were won by those who could make products affordable for the masses, not just by the inventors.
3. Open Source: From a “Necessity” to a Strategic Tool
Open source (where developers can download and modify models) was initially something U.S. companies were reluctant to adopt, as it could weaken their commercial advantage. However, due to limited computing power and weaker global brands, Chinese companies had no choice but to use open-source approaches:
- Open-source models quickly attract developers and overseas markets, helping to build ecosystems. For example, DeepSeek and Tongyi Qianwen’s models have gained many users internationally, even attracting American developers.
- Now the U.S. is also concerned: In 2026, more than twenty U.S. tech companies called for no restrictions on open-source models, fearing that China might use them to dominate the global market. What was once a necessity has become a strategic advantage for China.
The essence of open source is using an ecosystem to bridge the technological gap—you may have the most advanced models, but I can have the largest number of developers; who will have more influence remains to be seen.
4. Talent: China “Produces,” the U.S. “Gathers”
AI requires talents in mathematics, computer science, and engineering, which are areas that Chinese education has focused on for decades. Data shows that 38% of the world’s top AI researchers hold degrees from Chinese universities, yet 59% work in the U.S. Why?
- China as a Talent Producer: Mathematics, physics, and chemistry are strong subjects in Chinese education, and parents encourage students to pursue science and engineering, resulting in a large pool of foundational talents.
- The U.S. as a Talent Magnet: The U.S. offers a better innovation environment with capital willing to invest in long-term projects and laboratories that support original research, giving talent the belief that great things can happen there. Chinese innovators like Huang Renxun (NVIDIA) and Su Zifeng (AMD) achieved their success in the U.S.
China’s goal is not just to catch up with the U.S. in terms of technology but also to create an environment where talents are willing to stay and innovate.
5. The U.S. Aims for AGI, China Focuses on Productivity
The goals of AI development in China and the U.S. differ significantly:
- The U.S. Aims for AGI: Leading companies constantly discuss when machines will surpass humans and how superintelligence can be controlled, investing heavily even if it sounds like science fiction. They believe that those who achieve AGI first will wield significant power.
- China Focuses on Productivity: The government and companies are more concerned with how AI can upgrade manufacturing, transform supply chains, and create jobs. For example, using AI for quality control in factories, optimizing logistics routes, and aiding in medical diagnoses—these are practical applications with immediate benefits.
These differences stem from their respective resource endowments: the U.S. has the world’s strongest capital and software industry, allowing it to invest in the future; China, facing growth challenges and an aging population, needs AI to solve real-world problems quickly.
In conclusion, the article emphasizes that although the paths of AI development in China and the U.S. differ, the ultimate direction depends on what each country values—whether to pursue technological excellence or to ensure that AI benefits more people. This is more important than chips and algorithms alone.
This analysis explains the key differences between Chinese and American AI strategies in simple language, making it understandable for non-experts. The real question is not about who will “win” but who can turn AI into a force that truly changes people’s lives.