The US government’s decision to allow Nvidia to export its H200 AI chips to China has reopened a major channel for American semiconductor revenue, but whether Beijing will embrace the opportunity remains uncertain. In December 2025, President Donald Trump announced that the US would approve sales of Nvidia’s H200 to select Chinese customers under a strict licensing system, with a 25% revenue levy imposed on transactions.

Taiwanese manufacturers are positioning themselves for defense-sector expansion as the government prepares a NT$1.25 trillion (approx. US$40 billion) special budget to accelerate domestic weapons production and strengthen supply chain autonomy. The initiative aligns with a global surge in military spending and is expected to create new opportunities for unmanned systems, aerospace maintenance, and satellite communications.

US regulators have introduced a new inspection protocol for Nvidia’s H200 accelerators destined for China, requiring chips manufactured by TSMC to be shipped to the US for national security review before they can be re-exported. The unusual routing underscores Washington’s effort to maintain visibility over advanced computing hardware as global memory shortages threaten to slow new data center deployments through 2027.

PCB and semiconductor equipment manufacturer Ta Liang Technology reported a surge in orders as AI chip and AI server demand solidifies. Customers have been investing to expand capacity, driving demand for high-end CCD back-drilling machines, semiconductor advanced packaging measurement, and automated optical inspection (AOI) equipment to increase in 2025. This pushed cumulative revenue for the first 11 months of 2025 beyond the total for 2024, setting a record.