It has often been noted that China is in a position to overtake the US technologically. This is neither entirely correct nor entirely false. Despite the significant gains it has made in recent decades, China is still heavily dependent on software and hardware supply chains that are dominated by the USA.
AI development is an important illustrative case study. China’s key policies, such as the New Generation AI Development Plan (AIDP), are strongly reliant on leveraging access to US hardware and hardware manufacturing apparatuses in order to develop domestic applications. With secure access to these capabilities, China can rely on its strengths which, in the case of AI, are access to much greater amounts of data and the ability to rely on large numbers of relatively cheap but highly educated technical staff.
As AI in its current form is largely dependent on training algorithms on large data sets, China’s access to data is a massive advantage that can allow it to more efficiently and reliably train AI algorithms to specific purposes.
However, the US retains the ability to control and even restrict China’s supply of key hardware. AI development is currently reliant on the use of Integrated Circuits in the form of Graphical Processing Units (GPUs), Field Programmable Gate Arrays (FPGAs) and newer varieties of specialty AI Chips that are used for neural networks and deep learning. Industry watchers are widely in agreement that despite gains in recent years, Chinese companies are still years behind US-based industry leaders like NVIDIA and AMD on all of these formats in terms of the sophistication of their designs.
These challenges are compounded by the fact that, even for technologies that Chinese companies have caught up with the west in being able to design, it is still in most cases dependent on US-provided manufacturing processes to produce them. The US move on May 18th to ban Huawei from the use of semiconductor production equipment or software produced in the US has been posited by some observers to be a potentially mortal blow to Huawei’s prospects as a player in the international 5G arena – and, indeed, as a company in general. That this is a sanction that could be extended to any Chinese company and heavily handicaps the ability of China and its national champions to maneuver.
As we can see in the context of the wargame, should the US choose to directly target China’s key supply lines in this context, it is able to effectively arrest China’s ambitions, despite Chinese advantages and state support. This is something China is currently not ignorant of – hence the rise in investment in Integrated Circuit production facilities and technologies, which are intended to allow China to eventually bypass this strategic vulnerability.
Such an approach is not fully advantageous to the US, however. Cutting China off from the US economy would constitute a massive blow to the bottom lines of many American companies, which would mount vigorous political counteroffensives to prevent further escalation. As has been discussed, Apple earned between 15 and 20 percent of its revenue from Chinese sales in 2018-2019. Removing Chinese business from the equation to starve it technologically would result in catastrophic losses to American markets and businesses as well.
American companies that provide semiconductor equipment and tools also stand to lose significant amounts of revenue from the loss of Chinese clientele. This could, in turn, lead to them seeking to move operations out of the US, or even potentially being less able to compete on international markets with the loss of investment necessary for business or technological development.
Furthermore, in the long term, it is noted that China’s large domestic talent pool and the gains it has already made in developing vital technologies would mean that the use of a tactic along the lines of what the wargame proposes would only delay the development of China’s high-tech sectors, instead of fully crippling it.
The expected outcome in such a scenario would, unfortunately, result in fully bifurcated global standards for technology, which would result in the death of interoperability across specific geographical regions. Standards in place in the US may not be compatible with hardware and software in place in China, and devices used in either territory may be entirely unusable in another.
