What is the state of technological decoupling between the United States and China? Is there any cooperation between them in the tech sector?
One starting point is that it's very hard to measure decoupling because of how multifaceted the US-China scientific relationship really is. One way to measure decoupling would be initiatives and plans that have been released to increase the self-reliance of one's scientific and technological ecosystem from another or to break apart ties. Certainly, we have seen some of those actions over the past couple years from both the US and China. Another way to measure it could be the degree of co-authored publications that include one co-author from China and one co-author from the US. In those respects, the US-China relationship is still as strong as ever, and there might not be as much decoupling that's occurring. Other people look at how often patents in one country cite the patents of another country. In other words, what are citation patterns looking like among patents between the US and China. Some evidence has shown that there has been a decrease over the past couple of years in these cross citations for patents, and this could even be traced back to China’s accession to the WTO over two decades ago.
Part of the difficulty of understanding decoupling is we oftentimes look at only the most visible areas. Semiconductors comes to mind as an area where there have been huge actions and incentives from both sides to leverage that domain as an area for decoupling. One of the reasons why decoupling is so salient in that sector is because it's not only a crucial input to so many strategic areas like AI, but it's also controlled by a small amount of players and supply is concentrated in just a small amount of firms. Therefore, it's easier to leverage that as an area for one country to exert influence over another, and that's why you might be seeing more actions towards decoupling in that area. Whereas in other areas that where it's less concentrated, so maybe a field like smartphones, or even older industries, like electricity, there, there might not be as much decoupling occurring.
Can you speak a little bit more about any existing levels of cooperation between the US and China whether in the co-authored papers space, or more broadly on an industry or government level?
As someone who looks primarily at AI development and the technological domain of AI, when I've looked at the bibliometrics for citation patterns in the AI domain, the US and China actually cooperate with each other the most. When you map out the networks by which publications in the AI space are co-authored, and those co authorship networks more broadly, the US and China stand out in terms of the ties between the two countries. At the very least, the academic communities are still very connected, particularly in the AI space.
How has US-China technological decoupling impacted China given its reliance, particularly when developing AI, on semiconductor chips developed by US-based firms?
This is an interesting area where there are new developments every day. Recently, the US blocked shipments of high-end GPUs and AI accelerators that are made by the companies Nvidia and AMD. Chinese firms and in particular, Chinese AI firms, are dependent on these semiconductors to train large AI models. That was an effort by the US to slow down China's development of AI, particularly due to concerns about the Chinese military being able to build AI systems using some of these chips. The decoupling has definitely impacted China as China is largely dependent on the US for some of these high-end AI chips, and that's also why you're seeing so many efforts by China to develop its own homegrown companies that can make some of these high end chips that are used to train AI algorithms.
What are the costs of tech decoupling for the U.S.? Which sectors and firms in the U.S. have been adversely affected?
One of the obvious candidates is the semiconductor industry because they derive so many of their sales from China. These actions to try to decouple the supply chain away from China are definitely going to hurt the US, which is why we've seen the introduction and passage of the Chips and Science Act that is aimed at revitalizing the US semiconductor industry and improving US semiconductor manufacturing capabilities. That's not just an investment for the future, it's also a way to pay back the cost that the semiconductor industry has eaten in terms of decoupling efforts.
Aside from in the semiconductor space, are there other areas or industries that have been adversely effects by the technological decoupling?
The question with decoupling is always “it's going to hurt the US, but is it going to hurt China more?” It’s also a question of what kind of decoupling we really want. Obviously, everybody would be better off if globalization is maintained, but policymakers in the US have made the calculation that hurting China more is worth it at the cost of hurting our own companies. From a broader perspective, decoupling is always going to impose costs on US firms because US firms benefit from access to global innovation that works so they benefit from being able to sell products to Chinese consumers, they benefit from being able to collaborate with Chinese partners. All the costs are the essentially the opportunity cost of not having that extra globalization.
How much influence and power do multi-national firms, such as Google and Microsoft, have in advancing your idea of technoglobalism as an alternative to the current, seemingly more technonational order of world affairs? Have American tech firms tried to push back against tech decoupling?
From the perspective of some of these multinational firms, one point of contention has been whether these multinational firms like the Googles and the Microsofts of the world should be able to set up R&D centers in China. The same logic applies in terms of why some policymakers are arguing that they should pull out of China. Yes, these R&D centers are benefiting Google and Microsoft because they're able to tap into global networks of talent and they're able to understand what's happening in the Chinese AI ecosystem and benefit from technology transfers or just new ideas. But advocates of decoupling would say “okay, it does help the US, but it might help China more in terms of building up China's AI ecosystem.” This is the tension right now between technoglobalism, which is an approach that emphasizes that technology can't really be contained across borders and even these multinational firms are collaborating with other firms and they're not just beholden to their home nation anymore. In terms of the pushback, the argument that American tech firms have made is that the ability to tap into these global networks is crucial to US competitiveness, and them not being able to attract Chinese talent to the US or attract Chinese talent to work for multinational corporations in R&D labs, would hurt US competitiveness in the long run.
Have the actions and positions of these US-based tech firms that are trying to take a more technoglobalist approach been successful in pushing back against this decoupling?
At the very least, companies like Google are still setting up R&D labs across a variety of domains in China. Tesla, for example, is setting up a factory for electric vehicles in Shanghai, and that factory, I think, accounts for half of Tesla's global sales right now. There's a lot of examples by which these large companies are almost going to be forced to still operate in China. If the US pulls out of the entire Chinese innovation network, European firms will just step right in. A similar argument is oftentimes made with respect to college and universities, that is, if US universities closed off to Chinese talent, Chinese students would just go to some of the best universities in Europe. That again speaks to the fact that we have to understand this is a globalized system, and we have to understand it's not just a two-player game between the US and China, which oftentimes becomes the mindset when you talk about decoupling.
How would you evaluate China’s response to the tech decoupling between the U.S. and China? What are their successes? What are their mistakes?
One thing that should be pointed out is that US efforts to decouple are not just driven by economics or national security concerns. The Entity List that places certain companies on a blacklist for accessing US origin semiconductor equipment and chips, that list implicates countries that are allegedly involved in human rights abuses in Xinjiang. In terms of China's response, I think it's hard to say success or failure in that context for China. Getting off that entity list would involve acknowledging human rights abuses in Xinjiang and stopping them. That has not happened and in terms of China's response to tech decoupling to build up their own independent capacity, we are seeing limited success in semiconductor chips with SMIC being able to significantly increase their production of certain semiconductor chips, and that's been helped by a large government fund in integrated circuits. In terms of mistakes, I would say China thinking that it can go at it alone, in terms of producing all the critical key components for emerging technologies. No one country can go at it alone. The best approach is actually to diversify your sources of supply so that no one country can completely cut you off. You can survive a supply cut off by having a diversity of trade relationships and access relationships to key components.
Is it part of the CPC’s and Xi Jinping’s national plans to try and create an entirely domestic supply chain for these advanced technologies?
Yes. The Made in China 2025 plan, and even dating back to the medium, long-term plan for science and technology development, MLP, which was promulgated in 2006, all of those plans emphasize indigenous innovation and self-reliance. The Made in China 2025 plan sets out that 70% of components across some of these key strategic industries that the government has identified will be sourced from local producers as a benchmark. There is definitely a drive, of course, and there's debates and over how much self-reliance are needed for China to be secure. Oftentimes, these self-sufficiency targets are not met just because of the nature of the globalized economy, but we've seen it laid out in writing in some of these plans.
What would the future of the global technology landscape look like as a result of U.S.-China tech decoupling?
One potential vision of the future is, among industries that are deemed strategic or sensitive to sell to China, you have one supply chain, and maybe that supply chain is primarily located in China, so the Chinese government feels secure in that supply chain. And then you have another supply chain that is primarily sourced from the US and US allied countries, like South Korea and Japan. Therefore, US-China tech decoupling and the fact that these two ecosystems have elevated security concerns in certain industries, could create a global technology landscape that's more bifurcated among a US plus supply chain versus a China dominant supply chain. Another vision is that we have fits and bursts of decoupling when there's enough momentum around a certain industry, but overall, in a lot of the non-commonly covered industries that don't make the headlines every day like semiconductors and a lot of software-based industries where open source information flows are prioritized in a lot of these other industries, we don't see significant decoupling, and the global technology landscape looks pretty similar to what it has looked like over the past 10 years before the decoupling narrative gained momentum.
U.S. Department of AgricultureUSDA/LANCE CHEUNG/Photographer/USDA Photo by Lance Cheung, Public domain, via Wikimedia Commons