- US export controls briefly targeted AI software design tools essential for advanced chip development in 2025, before being rescinded weeks later as part of a trade agreement.
- Executive Order 14105 restricts American investors from funding Chinese AI, quantum computing, and semiconductor companies.
- Intel received approximately $7.86 billion, TSMC $6.6 billion, and Samsung approximately $4.745 billion in CHIPS Act grants.
US export controls have repeatedly reached into the AI software stack itself, targeting the design tools that feed advanced chip development, most notably in May 2025, when the Commerce Department briefly required licenses for electronic design automation (EDA) software sales to China, before rescinding the restriction weeks later as part of a trade agreement. The episode illustrates how far the sanctions architecture that began with GPU restrictions in October 2022 now extends, and how quickly it can shift with the broader trade relationship.
Chip Blockades and Beijing’s Response
Semiconductor export controls remain the sharpest instrument in Washington’s toolkit. The October 2022 restrictions, and their October 2023 updates, blocked exports of advanced AI chips including Nvidia’s A100 and H100 GPUs, along with the manufacturing equipment needed to produce them. The foreign direct product rule extends that reach to chips made outside the US if they rely on American tools or software, which covers most of the world’s leading-edge production.
Huawei is the clearest example of how this plays out. Procurement of high-end chips for its Ascend series has been severely restricted, forcing the company to intensify work on domestic alternatives. State-backed investment in foundries like SMIC has followed, alongside progress in less advanced process nodes. The brief 2025 EDA software restrictions demonstrated a further point of leverage: even if China builds the fabs, the design tools feeding them can be cut off by Washington on short notice — though the episode also showed how quickly such controls can be reversed when they collide with other trade priorities, as they were within six weeks.
Capital Curbs Reshape Venture Funding
Executive Order 14105, signed in August 2023 and now in phased implementation by the US Treasury Department, restricts American investors from funding Chinese companies working on advanced AI, quantum computing and semiconductors. Some investments require notification; others face outright prohibition. US venture capital and private equity firms that historically committed significant capital to China’s tech sector are now restructuring their portfolios accordingly, exiting certain ventures or declining to enter new ones. The practical effect is a narrowing of the funding pool for Chinese AI startups, pushing them toward domestic capital sources and widening the gap between the two countries’ innovation pipelines.
What the CHIPS Act Actually Bought
The grants have been substantial: Intel received approximately $7.86 billion, TSMC secured $6.6 billion, and Samsung was awarded approximately $4.745 billion for their US-based fabs under the CHIPS Act.
China’s Data Rules Create a Closed AI Environment
China’s data governance framework shapes AI development as much as hardware restrictions do, but from the inside. The Cybersecurity Law (2017), Data Security Law (2021), Personal Information Protection Law (2021) and the Interim Measures for Generative AI Services (August 2023) collectively impose data localisation requirements, cross-border transfer restrictions and algorithmic transparency obligations, all enforced by the Cyberspace Administration of China. For foreign AI companies, that means storing data locally, clearing security assessments before moving it out and often building separate, localised product versions for the Chinese market. The result is a bifurcated global AI market where the same model frequently cannot operate identically on both sides of the border.
The Talent Divide
Immigration policy is adding a research dimension to the hardware and capital fight. Increased visa scrutiny on Chinese STEM researchers and export control restrictions on academic collaboration in sensitive fields have made it harder for Chinese students and scholars to pursue advanced AI research in the US. China has responded by raising funding for domestic research institutions and offering substantial incentives to returning scientists. The longer-term effect is a gradual separation of research communities that have historically shared findings freely, though how far that separation extends remains unclear given the volume of open academic publication on both sides.
