A New Era of Semiconductor Geopolitics
The US-China semiconductor conflict has entered a new and more complex phase in 2026. What began as targeted export controls on advanced AI chips has evolved into a structural realignment of the global semiconductor industry, with both sides building independent supply chains, architectures, and increasingly incompatible AI infrastructure. The events of the past twelve months have made one thing clear: this is not a temporary trade dispute but a permanent reshaping of the $600 billion global semiconductor market.
The H200 Reversal: A Policy Zigzag
The most dramatic policy swing of early 2026 involves Nvidia's H200 GPU. In December 2025, President Trump announced that Nvidia would be allowed to sell H200 chips to China — reversing an earlier ban. Then a 25% tariff was imposed. Then a new licensing framework was created that experts immediately called contradictory.
The H200 reversal reflects the tension between national security objectives and economic interests. Nvidia's China revenue represents a significant portion of its data center business, and the company had engineered the H200 specifically to comply with export control thresholds. The policy zigzag has created uncertainty for both US chipmakers and Chinese customers, with Nvidia writing down $4.5 billion in inventory and warning investors of significant China revenue losses.
The BIS (Bureau of Industry and Security) has since shifted its approach from blanket denial to case-by-case evaluation of export license applications for chips like the H200 and AMD's MI325X — provided exporters meet a set of newly established security and compliance conditions. This creates a more dynamic and documentation-heavy compliance environment.
The Validated End-User System Expires
One of the most consequential regulatory changes of 2026 was the expiration of the Validated End-User (VEU) system on December 31, 2025. For years, select foreign chipmakers operating fabs in China — TSMC, Samsung, and SK Hynix — operated under special exemptions that allowed them to receive US-controlled semiconductor manufacturing equipment without seeking individual export licenses for every shipment.
Starting January 1, 2026, each company had to apply for and receive new annual export licenses from the US Department of Commerce to keep their Chinese fab operations running. All three received approvals, but the shift from automatic exemptions to annual renewals is significant: Washington now holds a renewal card it can play each year, and the terms can change with every renewal cycle.
TSMC's Nanjing facility, which produces 16nm and other mature-node chips, secured its license just as the new year began. Samsung and SK Hynix, whose China operations are central to global DRAM and NAND output, followed shortly after. The practical impact for now is minimal, but the structural leverage Washington has acquired is significant and lasting.
BIS Enforcement Surge
The BIS opened 2026 with an enforcement surge targeting chip exports to China, turning one long-term investigation after another into high-price penalties. Almost every case involved shipments to China's premier semiconductor foundry, SMIC (Semiconductor Manufacturing International Corporation), a company added to the Entity List in 2020.
In January, the BIS fined Germany's Exyte Management $1.5 million for its Chinese subsidiary's dealings with SMIC. In February, Silicon Valley-based Applied Materials and its South Korean subsidiary settled with BIS for $252 million — the agency's second-largest penalty ever. Further settlements with Solventum Corp. and Coastal PVA Technology followed in March and April.
This enforcement surge sends a clear signal: even as the US engages in trade negotiations with China, export control rules remain in full effect. As one compliance expert noted, the message is that the US "is going to continue to enforce these rules" regardless of the diplomatic climate.
China's Response: Domestic Acceleration
China has responded to export restrictions with an accelerated push for semiconductor self-sufficiency. The country's 2030 goal of 80% chip self-sufficiency has driven massive investment in domestic foundry capacity, with SMIC allocating $7.5 billion in capex for 2026. Chinese AI chip startups including Moore Threads, MetaX, and Biren have raised billions to develop domestic GPU alternatives.
However, the actual impact of export controls on China's AI capabilities is nuanced. While restrictions have clearly hampered China's ability to manufacture advanced chips at scale, they have not prevented Chinese AI companies from producing competitive models. Firms like Alibaba, DeepSeek, and ByteDance have developed large language models that score competitively on established benchmarks, often using creative workarounds to access limited compute resources.
DeepSeek founder Liang Wenfeng has publicly stated that "money has never been the problem for us; bans on shipments of advanced chips are the problem." This sentiment reflects a widespread view among Chinese AI leaders that export controls are effective in constraining their capabilities.
The MATCH Act and Allied Alignment
In April 2026, the MATCH Act was introduced in the US House, calling on allies to align with expanded export controls. The bipartisan bill reflects congressional concern that allied countries — particularly the Netherlands (ASML) and Japan (Tokyo Electron, Screen) — could become channels for controlled technology to reach China.
The US-Japan-Netherlands trilateral framework on semiconductor equipment exports has been a key pillar of the control regime, but its effectiveness depends on continued alignment among the three countries. Japan has implemented its own 23-item export control list covering semiconductor manufacturing equipment, and the Netherlands has maintained restrictions on ASML's advanced lithography systems.
The Bifurcation of Global Supply Chains
The long-term consequence of export controls is the bifurcation of global semiconductor supply chains. Two distinct ecosystems are emerging: a Western-aligned ecosystem centered on US, European, Japanese, and Korean technology, and a China-aligned ecosystem built on domestic equipment, materials, and manufacturing capabilities.
This bifurcation has significant economic implications. Dual supply chains increase costs for everyone, as duplicate R&D investments and parallel manufacturing facilities are required. It also reduces the benefits of global specialization that drove semiconductor cost reductions for decades.
For chip designers, the bifurcation creates a difficult strategic choice: which ecosystem to align with, and how to manage the risks of being locked out of one market or the other. Some companies are pursuing "geopolitically neutral" strategies, while others are making explicit bets on one side.
The Outlook
As of mid-2026, the BIS is unlikely to introduce major new export control rules amid ongoing US-China trade negotiations. Instead, the focus is on strengthened enforcement of existing rules — which gives the US flexibility to escalate or de-escalate based on broader diplomatic dynamics.
The long-term trajectory remains uncertain but the direction is clear: semiconductor manufacturing is being reshaped by national security priorities, and the era of purely economic decision-making in chip supply chains is over. For the foreseeable future, geopolitics will be as important as technology in determining where chips are made, by whom, and for whom.