The restrictions attempt to constrain Beijing’s emerging technology development in two ways. First, they limit China’s access to advanced semiconductors necessary for data-intensive artificial intelligence (AI) models, supercomputers and hypersonic missiles. Second, the measures curtail China’s ability to create these cutting-edge chips by preventing China from using US-built semiconductor design and manufacturing equipment, including the electronic design automation (EDA) software required to design advanced semiconductors.

But engineers at China’s top technology university — Tsinghua University — indicate that the 7 October export controls may not fully accomplish Washington’s objectives and that the sanctions’ scope should expand.

Conversations with over 60 engineers on Tsinghua’s campus reveal the prevalence of continuous EDA software limitation workarounds and the downsides of only targeting the most advanced-node chips, as opposed to so-called ‘mature node’ chips that offer less density of processing power for a given physical size.

Several students at the School of Integrated Circuits at Tsinghua (a ‘chip college’ established in 2020 to foster China’s top semiconductor talent) noted that they can easily circumvent EDA restrictions. Despite sanctions, students use Cadence and Synopsys — two of the most popular EDA providers from the United States — through backgate channels or special licenses. They then send advanced chips they design with EDA software to the Taiwan Semiconductor Manufacturing Company to manufacture — achieving the entire process Washington aims to restrict.

While Chinese firms have released their own EDA software, most students still rely on US technology. One student explained that ‘it’s custom’ and that ‘we can’t just start using new software, the entire process is connected. No one is willing to use it’. Another student noted feeling ‘panicked under the pressure’ of future sanctions and worried that ‘if the United States one day decides to stop letting us use EDA software, we’ll go blind. I’m terrified about that’.

Tsinghua engineers that access US EDA software may conduct dual-use or military end-use research for the PRC’s military-civil fusion initiative. The Australian Strategic Policy Institute ranked Tsinghua a ‘high-risk’ university for its ‘range of military research’, top-secret clearance for classified defence research and direct military funding. Tsinghua’s defence research sectors include navigation technology, AI, and air-to-air missiles — all technologies that rely on advanced semiconductors.

Even if Washington could successfully limit Beijing’s cutting-edge semiconductor development, this restriction alone would not be enough. Most military technology, including cutting-edge space technology, relies on less sophisticated chips that can handle extreme environmental changes and temperature shifts. Degrading China’s national security and defence industries requires limiting these ‘mature node’ chips.

A PhD student set to graduate in 2024 and ‘manage military technology development’ confirmed the current sanctions’ inability to enfeeble the People’s Liberation Army (PLA). He describes the October chip ban as ‘ineffective in terms of military use’.

While the PLA relies on foreign-produced chips for cutting-edge research, the PhD student noted that this research is secondary. He said, ‘we could use AI for drones and coordination, but that’s more like 15–20 years down the line’, and that ‘right now, we are more focussed on the applications that need chips 28nm and above like missiles, space and jets’. A 2022 RAND Report confirms that military technology development — such as space technology, radio frequency communication, laser-based sensing, edge computing, and integrated silicon photonics — mostly relies on mature chips.

According to the Center for Strategic and International Studies, Washington did not target China’s mature chip (16nm and above) industry because this could have exacerbated inflation and angered foreign firms reliant on China’s older semiconductors. But the United States may regret this decision in the long term. Not only has this choice been a blessing for PLA technology development, but by targeting cutting-edge and not older semiconductors, Washington has stimulated Beijing’s mature chip industry.

This lopsided restriction increases China’s ability to dominate the global mature node market. China’s premier logic chip producer — Semiconductor Manufacturing International Corporation — plans to open new 28nm node production facilities in Shanghai, Shenzhen, Beijing and Tianjin over the next two years. Chinese strategists assert that mature node dominance provides Beijing with a future ‘bargaining position’ to limit US access to older semiconductors. Many Tsinghua engineers also noted China’s mature node market potential and 28nm chips’ role in China’s overall technology development.

Tsinghua’s robust network of top-tier engineers provides unique insight into the effects of US semiconductor sanctions. Yet uncertainty still exists on whether engineers can design cutting-edge chips on par with foreign firms. As technology rapidly advances, questions emerge regarding mature nodes’ ability to sustain Chinese military development, the future of China’s domestic semiconductor industry, and the role of US sanctions in moulding this environment.

If the United States intends to halt Chinese President Xi Jinping’s commitment to the autarkic mastery of critical technologies, the October 2022 export ban will not suffice. Washington’s ‘scalpel’ should not expand to a sledgehammer. But the United States should increase EDA software restrictions and recalculate the perceived trade-off between mature-node chip controls and inflation risks.

Semiconductors propel the entire technology industry. The nation that controls the global production network of this critical tool holds unprecedented geopolitical power. Given the PLA’s aggression in the region and Beijing’s unfair market practices, China occupying this position does not bode well for the Indo-Pacific — or the world.

Christina Knight holds a B.S. in Symbolic Systems Artificial Intelligence, a Master’s in Philosophy, and a
Master’s in Global Affairs from Stanford University and the Schwarzman Scholarship.