TLDRs;
- The US is weighing a plan to replace automatic export approvals with annual licenses for Samsung and SK Hynix factories in China.
- Washington aims to allow continued factory operations while blocking upgrades that could advance China’s semiconductor industry.
- Nvidia faces parallel restrictions as it negotiates with the White House to sell scaled-down AI chips in China.
- The policy reflects broader US-China chip tensions and could reshape global semiconductor supply chains and investment strategies.
The United States is considering replacing its current system of automatic export approvals for semiconductor supplies with annual review requirements, a move that could reshape how global chip giants like Samsung Electronics and SK Hynix operate their factories in China.
At present, both Samsung and SK Hynix enjoy “validated end user” (VEU) designations that allow them to ship equipment and materials to their Chinese operations without repeated bureaucratic hurdles. These authorizations, however, are due to expire at the end of 2025.
According to people familiar with ongoing discussions, Washington is now weighing whether to switch to a “site license” model, requiring the South Korean companies to apply each year for permission to send restricted chipmaking tools and materials.
Balancing supply chain stability and security risks
US officials say the new framework would strike a balance by allowing existing factories to maintain operations without sudden disruptions, while preventing upgrades or expansions that could strengthen China’s domestic semiconductor industry.
In effect, the measure would protect continuity of global supply chains while ensuring that cutting-edge chip technology does not flow freely into China.
The decision, however, has not been finalized. Negotiations with South Korean officials remain ongoing, reflecting the delicate balance Washington must maintain between securing critical technologies and honoring commitments to allies.
Nvidia’s parallel struggle highlights wider stakes
The debate comes as American chipmaker Nvidia is also in discussions with the US government over selling modified versions of its AI chips to China.
CEO Jensen Huang recently revealed that the company is seeking permission to offer scaled-down Blackwell chips, up to 50% less powerful than its flagship products, in order to access China’s estimated $50 billion AI market.
Nvidia’s situation underscores the stakes of Washington’s evolving semiconductor policy. Export restrictions aimed at curbing China’s access to advanced processors may also fuel Beijing’s determination to achieve technological independence. In fact, Chinese firms such as Cambricon have seen significant growth amid these limitations, while Beijing has earmarked hundreds of billions of dollars for domestic chip and AI development.
Long-term implications for global chip supply
Analysts warn that shifting from automatic approvals to annual licenses could increase uncertainty for multinational chipmakers. While operations in China would likely continue, firms may face rising costs, regulatory delays, and limits on investment planning.
More broadly, the policy highlights how semiconductors remain at the center of US-China tensions, an arena where economic competitiveness and national security converge.
For South Korea, which relies heavily on both the US for security and China as a manufacturing base, the decision could force difficult adjustments. If finalized, the annual approval model could redefine how its largest chipmakers manage production strategies in one of their most important markets.
