Can China’s Semiconductor Technology Advance Be Stopped?

By Jennifer Read, Editor, EMSNOW

Nearly four years after the US first imposed sweeping export controls on advanced chipmaking technology bound for China, the electronics manufacturing industry is still grappling with the same question: are these controls working?

The answer emerging from 2026 data is nuanced. Controls have clearly slowed China’s progress at the leading edge. They have not stopped it, and in some respects they’ve accelerated a longer-term shift that may prove harder to reverse.

The controls have teeth

The numbers on China’s advanced frontier capacity are stark. Industry analysis indicates China is on track in 2026 to domestically produce a total volume of AI chips equal to just 1-2% of US output [1]. Cut off from the most advanced lithography and design tools, Chinese fabs have been pushed toward older equipment and less efficient processes, which constrains both the quality and volume of what they can produce domestically.

That gap traces back to a real dependency problem. Before allied export controls tightened, domestically made semiconductor manufacturing equipment (SME) accounted for only about 10-15% of China’s domestic market [2], leaving Chinese chipmakers reliant on Western and allied toolmakers for the vast majority of their fabrication capability.

The regionalization drive is real and accelerating

Where the controls have arguably backfired is in incentive structure. Rather than accepting a permanent technology gap, China has redirected national resources toward supply-chain independence at a scale and urgency that predates the controls. Industry sourcing suggests Chinese chipmakers now regard escaping dependence on foreign suppliers as more strategically important than any near-term technological gain [2].

That shift shows up in the data. TrendForce projects China’s domestic share of its own AI chip market will reach 50% in 2026 [2], up sharply from a few years ago. On the chip-design side, Chinese firms are shipping at real volume: Alibaba has reportedly delivered over 100,000 units of its Zhenwu 810E AI chip, and at least nine Chinese AI chipmakers, including Huawei, Baidu, and Cambricon, have each exceeded 10,000 shipments or total orders [2].

China’s equipment strategy has run on two tracks simultaneously: large-scale stockpiling of foreign tools to keep wafer fabrication and packaging R&D uninterrupted, while domestic equipment makers pour capital into their own tool development [4]. According to TrendForce analyst Joanna Chiao, near-term breakthroughs at the leading edge remain unlikely, but the next five years should bring steady, incremental gains across the domestic supply chain, validating domestic equipment in live production, scaling up mature and legacy node capacity, and closing the efficiency and yield gap with foreign tools [4].

There have also been individual signals of ambition beyond incremental progress: Shanghai startup Dishan Technology said in April 2026 that it had made significant headway on 2 nm AI chip prototypes [4], though independent verification of that claim remains limited.

Where the line still holds

The most persistent chokepoint remains extreme ultraviolet (EUV) lithography — the technology required for the most advanced logic nodes. Despite years of investment and government-backed initiatives like the Shenzhen-based SiCarrier model (combining state funding, market-style operation, and global talent recruitment) [5], there’s no clear public evidence China has cracked EUV production at scale. Progress in wafer manufacturing and advanced packaging specifically has been hindered by international equipment controls even as China gains ground elsewhere in the supply chain, according to TrendForce’s Chiao [4].

Even among the countries imposing controls, there’s no unified strategy. US enforcement has shifted from loud escalation toward quieter, more selective application [6], even as the administration approved export of Nvidia’s H200 and AMD’s Instinct MI308 GPUs to China in 2026 [2]. That approval triggered pushback from congressional hawks: the House Foreign Affairs Committee advanced legislation that would shift AI chip export license authority from the Commerce Department to Congress, and the House Select Committee on China has pressed the Commerce Secretary to restrict China’s access to advanced chipmaking equipment [6].

Meanwhile, China has developed its own leverage. In 2026 it widened rare-earth export controls to cover 10 US firms and 14 EU firms [7]. These are the materials that feed directly into semiconductor manufacturing equipment, servers, and the actuators, motors, and magnets found throughout electronics supply chains [7]. Notably, actual shipment volumes haven’t moved in lockstep with the headline restrictions; rare-earth magnet exports actually rose from June to July 2026 even as the control list expanded [7], underscoring how complicated enforcement and compliance have become on both sides.

Export controls have measurably slowed China at the leading edge and preserved a real capability gap, but they have not stopped China’s broader semiconductor ambitions, and they appear to have hardened Beijing’s resolve to build an independent, if currently less advanced, domestic supply chain.