Since late January, the hottest event in the global technology field is the rise of DeepSeek, a Chinese artificial intelligence company. The two AI products released by this previously unknown company shocked the industry. Its downgraded chips and low training costs have achieved results close to or reached the level achieved by OpenAI with the most advanced chips and large investments, which seems to subvert people’s perception of the AI development path built by the United States. In fact, the shock brought by DeepSeek has gone beyond the technical level and has become one of the hottest topics in global public opinion during this period. Many people have praised it with exaggerated language, especially in China. Some people in the science and technology community even described it as a “phenomenal” event, raising its significance to the height of changing China’s “national destiny”.
Whether DeepSeek is a major breakthrough in China’s development of artificial intelligence, or even a challenge to the hegemony of the United States in the field of AI, it may take a longer time to make an accurate judgment. However, the shock wave it has brought cannot be underestimated because of the value stance of China. On the contrary, it should be taken seriously because it comes from China, which is under the technological blockade of the United States. The success or failure of the competition between the United States and China essentially depends on the competition between the two countries in the field of science and technology. At present, the public opinion on DeepSeek focuses on the performance and technology of the product itself, but the technological competition between the United States and China involves deep-seated issues related to science and technology and industrial policies, education system, technology development path and development model. Therefore, the reflection on the shock wave of DeepSeek cannot be limited to itself. It is necessary to jump out of its technical level and focus on at least four aspects: whether China’s technological development route will be consolidated and strengthened, whether original innovation can break through under the current system in China, whether the training of Chinese AI talents will increase significantly, and whether the US technological blockade of China under the open source model is still effective.
DeepSeek has given the Chinese authorities confidence in the development of science and technology
. On the surface, the rise of DeepSeek is not closely related to China’s path of science and technology development, because it is a private equity company that transformed from quantitative investment. There is no evidence that it has received official policy and financial support, but in fact, from DeepSeek, we can see the shadow of China’s path of science and technology development. The development path of China’s science and technology since the reform can be briefly summarized as four stages: “imitation-following-parallel-partial leading”. In the early days, China was almost completely dependent on the West in terms of technology. Through “market for technology” and introduction and digestion, it completed the accumulation of technology. This process lasted for at least 20 years, and large-scale technology imitation was basically ended at the beginning of this century. However, in the later stage of the imitation stage, with the gradual improvement of China’s technological capabilities, it also began to follow the strategy of following the pace of technological development of the United States and Europe in cutting-edge technologies, major projects, and key industries, in order to quickly shorten the technological gap with the West.
Around 2015, China entered the stage of paralleling with the West from following the West. Parallel is what the Chinese government calls independent innovation. The strategy of independent innovation in science and technology was actually proposed in the later period of Hu Jintao, but the “innovation” at that time was more about following Western technology under the banner of “innovation”, and was still trapped in the path dependence of “tracking hot spots” as a whole. This situation did not change until 2015. China increased its investment in research and development of science and technology and consciously promoted the innovation strategy. The most prominent manifestation was “Made in China 2025”. After the trade war with the United States, it was impossible not to innovate independently due to the technological blockade of the United States and its allies.
Overall, China’s current technological imitation and following stage has passed, and it is in the stage of partial following and parallel development. In addition, there has also been a partial leading situation in the fields of electric vehicles, quantum communications, supercomputing, etc. However, it must be pointed out that China’s current scientific and technological innovation and partial leading are still mainly engineering technology innovation and leading, rather than original innovation as people often say. Therefore, the emergence of DeepSeek may indeed greatly enhance the confidence of the Chinese authorities in the path of scientific and technological development, making the authorities believe that as long as they firmly follow this path and make good use of the new national system, China will become a scientific and technological power and break through the US technological blockade.
However, this so-called independent innovation path is not always effective. In the past few years, in order to break the technological blockade of the United States, the authorities have relied on special fiscal investment or state-owned enterprise industry funds to strongly support many companies. The resources invested are amazing, but the results are far from satisfactory. For example, the national artificial intelligence key projects led by many academicians have produced little scientific research results in the past few years. Scientific and technological innovation cannot be without the government, but the way the government appears will determine the success or failure of innovation. China’s problem lies precisely in the fact that the government plays an overly active role and personally participates in planning and operation. DeepSeek’s success was not achieved under the care of government funds. The authorities may only see its current achievements, but may not be able to realize this.
DeepSeek does not represent a breakthrough in China’s original innovation.
Closely related to scientific and technological innovation is the question of whether China is capable of original innovation, which is a controversial topic. Most people will admit that China’s gap with the West, especially the United States, in terms of original technology is not in the same order of magnitude. Since the modern scientific and technological revolution, China has almost no major scientific and technological and basic theoretical innovations in the history of science and technology. So far, only one local scientist has won the Nobel Prize. These two aspects show the lack of original technology in China. Even if a person is patriotic, he cannot deny this fact.
So, can DeepSeek be regarded as a breakthrough in China’s original innovation? The answer is no. Although some supporters regard it as a “milestone in China’s original innovation”, even DeepSeek itself does not think so. If you ask it this question, its answer is that its technical breakthroughs such as reasoning ability are more of an improvement at the application level, rather than original innovation in basic theory, and do not represent a breakthrough in original innovation. China’s
long-term backwardness in original innovation has led to a serious lack of confidence in innovation among scientific and technological talents. But this is probably not the main factor. There are two main reasons: one is the clamping of freedom of thought by the authoritarian system. Original innovation requires a space for free discussion, and the results of innovation are also likely to affect the interests of the rulers. This is why China has not had any original scientific and technological achievements in the past few hundred years. Second, the development of industry and technology has lagged behind for a long time. Major breakthroughs or innovations in scientific and technological theories and technologies are the result of industrial and technological progress. If a country’s industry does not develop, the government and enterprises will not have more funds to invest in research and development, and there will be no technological progress. Scientific and technological innovation emerges in the process of technological progress to solve the bottlenecks encountered in industrial development.
At present, China’s industry has come up and has the so-called full industrial chain advantage, but the degree of authoritarianism of the regime is much more serious than most of the time before, which makes China’s scientific and technological innovation still limited to the level of technological breakthroughs such as 1-100. In the original innovation of 0-1, although scientific and technological talents are also working hard to seek breakthroughs, they have not achieved much so far. The emergence of DeepSeek can be seen as the result of China’s strong industry, which may mean a step closer to original innovation, but after all, it is not original innovation. To break through the limitations of original innovation, China still needs longer-term technological accumulation and a freer atmosphere of exploration.
However, one thing is certain, that is, the success of DeepSeek will stimulate more young Chinese people, especially college students, to invest in the field of AI. DeepSeek has a young entrepreneurial team, including its founder, who is not very old. Among the 140 members, 90% are fresh graduates and doctoral interns, with an average age of 26. Young people are relatively open-minded, curious, and not bound by so many rules and regulations. Now, this “youth army” has become a popular star in China because of DeepSeek, and some of them have been poached with an annual salary of tens of millions. From past experience, this is bound to set off a wave of AI fever among young people in China. The AI majors in universities may be overcrowded, and companies will also recruit more specialized talents to meet the needs of AI development. China does not lack engineering and technical talents. Every year, it trains more science and engineering students than the top few Western countries combined. More young people joining the AI industry is beneficial to the development of AI and further enhances China’s technological progress and national competitiveness.
On the other hand, it should also be noted that Chinese universities have not yet gotten rid of the rut of specialized education and paper education. The number of doctoral students produced is the highest in the world, but there are few available talents. Although most of Deepseek’s young entrepreneurial team are locally trained, its selection and employment standards are also a kind of subversion of China’s current education model and talent utilization model.
Helping the United States improve its technology control over China
From the perspective of the United States, the rise of DeepSeek also challenges whether the United States’ technology containment of China under the open source model is still effective. Public opinion has been asking whether the United States’ development model, which is highly dependent on computing power and based on the underlying logic and technical moat of “computing power is power”, is necessarily inevitable for the development of AI? If not, then under the open source model, can the United States’ technology containment of China still delay China’s AI development and widen the technological gap between the two countries?
The open source model is a decentralized technology ecosystem that gathers the wisdom of many people to seek technological breakthroughs and innovations. In terms of value, it seems to be naturally superior to the closed model. When Musk participated in the establishment of OpenAI, he advocated open source. After DeepSeek came out, OpenAI founder Altman seemed reluctant to admit that the closed source development idea was wrong. Therefore, trying to use centralized means to curb the decentralized ecology is like using a city wall to block a tsunami, which is theoretically difficult to succeed.
This does not deny that the United States’ technology control over China has been effective so far. Regarding this point, the emergence of DeepSeek, which is strongly promoted by Chinese official and public opinion, means the failure of the US technological containment. It is more of a propaganda strategy than the real situation. The reason is simple and clear. If there is no US control over China in chips and software, China’s overall AI development capabilities may be closer to the United States than it is now. Although the technological containment has forced China to “find another way”, the emergence of DeepSeek seems to have made the US sanctions fail. However, if the United States adopts more intensive technical control over China than before and completely isolates China’s AI development, whether DeepSeek’s open source model can still last, and whether further technological breakthroughs will encounter some insurmountable bottlenecks, it is not impossible.
The shock wave of DeepSeek has made the outside world not underestimate China’s technological capabilities. The United States does need to be vigilant about this, but it is unnecessary to exaggerate its ability to subvert the development path of existing AI technology. Its technological innovation is currently just an isolated case, and it does not indicate that China’s overall innovation environment has fundamentally changed. Therefore, its self-detonation is a good thing for the development of AI and the improvement of technology control strategies in the United States.