Mike Magee
Last week, Jensen Huang incited Trump's wholesale reversal of policy that prevented NVIDIA's H20 chip sales to China after visiting the White House. What did Jansen say?
Of course, we can only guess. But he may have shared the results of a proprietary report by AI researchers in digital science, which shows that corrections for instant policy courses are crucial. In addition to more than 50% of all AI researchers currently, their research also records that “in 2000, Chinese-based scholars produced only 671 AI papers, but in 2024, their 23,695 AI-related publications have produced more than the United States (6378) (6378) (6378) than the United Kingdom (6378), United Kingdom (2747), European Union (2747), European Union (10,0,055).
David Hook, CEO Digital Science It was declarative at the opening ceremony of the report, saying “The U.S. influence in AI research is declining and China is now dominant.”
China now supports about 30,000 AI researchers, compared with only 10,000 in the United States. That number is shrinking due to our tariff and visa pranks, as well as the government’s public attacks on our Prime Minister’s academic institutions.
Economics professors David Auto (MIT) and Gordon Hanson (Harvard) famously described the elements of “China's shock 1.0” with “their research on globalization, especially the rise of China,” reshape the U.S. labor market.” In 2013. It was “a single process – China’s transition from Maoist central planning to a market economy in the late 1970s, which quickly transferred the country’s labor and capital from collective rural farms to capitalist urban factories.”
As a result, between 1999 and 2007, a quarter of all manufacturing in the United States disappeared. Today, China's manufacturing workforce exceeds 100 million, bringing the number of jobs in the U.S. manufacturing industry to 13 million. These figures peaked when China's low-cost labor supply reached its peak a decade ago. But now, China has clearly been expecting, and the current government and its advisers are left behind in the rearview mirror.
Welcome to “China Shock 2.0” Auto and Hanson wrote in a recent New York Times editorial. But this time, their leaders focus on “the key technologies of the 21st century… (it) will continue to use China’s resources, patience and discipline to compete fiercely.”
The highly respected Australian Strategic Policy Institute funded by its Department of Defense has been tracking the number of published innovative technology research in the United States and China. They believe this is the measure of expert opinion that the greatest source of innovation. In 2007, we led China in 60 of the 64 “Border Technologies” to lead China in the first four years.
Twenty years later, the table has flipped, and China leads the U.S. in 57 of 57 of 64 categories.
In AI algorithm research, 29% of papers in China account for 12% of ours. In high performance computing, they have 36% to 13% of ours. In machine learning, they tilt the scale at 40% speed, while our 11% is.
According to Auto and Hanson, China is the “highest predator” in the battle with “economic Darwinism” and Chinese industrial policy. As Musk is demolishing our national planning and funding agencies, the Chinese government has been targeting our most innovative sectors including “Aviation, AI, telecommunications, microprocessors, robotics, nuclear and fusion capabilities, nuclear and quantum computing, quantum computing, biotechnology, biotechnology and pharmaceuticals, solar and turrets.”
The warning sounds harsh, but it is a misunderstanding of historical proportions. It all boils down to one word. tariff. There is no more hostile country than China when it comes to Trump’s tariffs, and this is Harvard Business Review’s condemnation of this intentional isolationism in January 2025: “Even if there are data privacy and security issues, the U.S. LLM will ignore the threat of the crisis of China’s LLM.”
In the paper, Harvard Business School professor Raj Choudhury worked with Natarajan Balasubramanian, professor of business administration at Syracuse University and Mingtao Xu, an economist at Tsinghua University, to answer the timely question: “Why shouldn't DeepSeek be surprised”.
For those unfamiliar with this controversy, the American Life AI Superegos has been running too much lately, with huge back pats and philosophicalization (between the Rocket Journey and Mars’ obsession), and the then-less-known Chinese price, DeepSeek, DeepSeek, made the Apple cart open smaller, more efficient, more effective, and worse, Apple’s carts make Apple carts more annoying. ”
The timing isn't even more awkward for our technocratic elites – anthropomorphic Claude, Google's Gemini, Open AI's Chatgpt, Microsoft Copilot, Musk's X.AI and Metaai.
These proprietary, self-proclaimed tech figures are totally committed to knocking down each other. But that doesn't stop them from embracing the atmosphere of open AI, Sam Altman, who recently wrote an op-ed titled “The Age of Intellect” in which he explained: “Technology takes us from the Stone Age to the Agriculture Age and then to the Industrial Age. Intellectual age Paved with calculations, energy and human will. ”
These American billionaires put their money in their mouths, burning a lot of capital when building a brand new industry – creating the generated AI factories that require so much energy that our once nascent nuclear industry investments have now passed the roof, and Musk has officially registered a new city for himself and his AI followers.
Meanwhile, Rand researchers like Kyle Chan report: “From chips and data centers to energy, China is applying state support across the entire AI Tech stack.” The Chinese government’s strategy focuses on three coordinate building blocks: 1) computing power, 2) vertically integrated and accessible Chinese data and 3) skilled engineers.
Meanwhile, the government has been funding a network of private labs dedicated to public/private collaborations and inspiring local government leaders in cities across the country to support hundreds of AI research startups. A city in southern China, Hangzhou already owns giant Alibaba, created part of the town called “Dream City” and is now the birthplace and permanent residence of DeepSeek.
Rand's Kyle Chan recently marked the fact that China has turned adversity (Trump's tariffs) into an incentive tool. By keeping its LLM discovery open source (as US companies tend to have proprietary controls to improve future profitability), authoritarian China (should be against transparency and accessibility), this is completely in the opposite direction of the U.S. technology industry, according to the Atlantic.
Why DeepSeek now? Management experts say this is just “a classic theory of destruction in the game.” A typical case is that decades ago, steel mills in the steel industry were serving high-end customers of high-end steel steel, exceeding inefficient arc furnaces. In response, old plants with inferior technology are customized and specifically targeted at low-end tasks such as creating steel rebars to enhance concrete structures. The incumbent was then forced to “suppose market share to disrupters.”
In this case, DeepSeek may be the inferior technology of the next greatest version of Chatgpt, but it seems to be enough to provide “industry-specific applications powered by its LLMs that have been deeply integrated into China’s digital ecosystem.” Chinese LLMs are more specific and less versatile. They require less computing power and price it on the lower end.
How low is it? For software developers around the world, DeepSeek-R1 is 95% cheaper than O1 products using open AI and is effective. Microsoft CEO Satya Nadella said that such pricing would cause AI to “sky, turning it into something we simply cannot get.”
Bottom line? Under the nose of the United States, China has attracted global software developers and users through open codes and brought the top U.S. competitors (at least for now) into the driver's seat, thus “democracy.”
Mike Magee, MD, is a medical historian and routine journalist at THCB. He is the author of Code Blue inside the American Medical Industry Complex. (Grove/2020).
