US chip controls were designed to slow China — semiconductor exports just rose 130%

China’s semiconductor exports rose 129.8% compared with the same month last year — the sharpest year-on-year jump since the current US export-control architecture came into force in 2022. The figure arrived alongside a broader trade surplus of $119.1 billion, up from $112.5 billion in July and tracking toward a full-year record. China’s General Administration of Customs published the data covering the full spectrum of exported chips: logic circuits, power management semiconductors, memory, and the automotive-grade components embedded in every major Chinese electric vehicle.

The scale of the semiconductor surge confronts a core assumption in US trade and technology policy. Export controls introduced by the Biden administration in 2022, expanded in 2023, and extended again under the current administration were designed to deny China access to the equipment needed to manufacture advanced chips — those below the 16-nanometer threshold required for AI training accelerators and next-generation processors. The stated logic: without frontier hardware, China’s technology sector would slow. The August numbers test that logic. Not by showing that China cracked the advanced-node problem — it has not, at least not at scale — but by showing that the tier the controls left largely untouched has become a formidable production machine in its own right.

China’s automobile exports climbed 43% year-on-year in August, with semiconductor-rich EV drivetrains driving much of the count. Exports to the US rose 34.4% to $42.5 billion in the month, widening the bilateral imbalance to around $29 billion. Southeast Asia absorbed a 30.2% jump in Chinese shipments; Latin America, a 17.5% increase. Each of those regions is a manufacturing and consumer ecosystem increasingly built on Chinese-supplied components.

The year-on-year comparison is not a clean read. August 2025 was a weak period for Chinese chip output — supply-chain constraints suppressed production — which makes the 2026 base look particularly strong in retrospect. The customs data also does not distinguish between chips manufactured domestically in China and components assembled in bonded processing zones from foreign-sourced wafers. Chi Lo of BNP Paribas called China “very competitive in its tech goods exports,” but analysts focused on the advanced end have noted that leading-edge production, specifically targeted by US controls, remains constrained by equipment access. The gap in the control architecture runs through the legacy tier — below the frontier, above the commodity — not at the top.

The workers and investors who feel this most directly are not in AI-chip fabs. They are in the industrial and automotive segments where Texas Instruments, Infineon Technologies, and STMicroelectronics compete on cost and delivery reliability. TSMC’s Arizona facility and Intel’s Ohio plant are building advanced-node production — not directly exposed to the same price pressure. But European automakers sourcing components across the semiconductor market are already competing against Chinese-designed chips in Southeast Asian showrooms. That is a consequence the aggregate trade figures only partially capture.

The next data point arrives in early October, when China publishes September trade figures. The Commerce Department’s next review of the Entity List — governing which Chinese firms can receive advanced chipmaking equipment — is expected before year-end. The August semiconductor numbers will enter that review carrying considerable weight.

Tags: China, trade policy, semiconductor exports, Texas Instruments, US chip controls