US-China AI Chip War 2026: Export Controls, Rare Earths & Decoupling

The US-China AI chip war has entered a decisive new phase in early 2026. On January 15, the Biden administration's Bureau of Industry and Security (BIS) published a final rule that fundamentally tightens export controls on advanced semiconductors destined for China and Macau. The new policy introduces stringent total processing power (TPP) thresholds and shifts licensing for key AI chips—including NVIDIA's H200 and AMD's MI325X—from a presumption of denial to a case-by-case review system. In parallel, Beijing has retaliated with its own export controls on rare earths and critical minerals, weaponizing its near-monopoly on processing to pressure Washington. This escalating technological decoupling is fragmenting global semiconductor supply chains, forcing allies to choose sides, and setting the stage for parallel AI ecosystems that could reshape innovation for the rest of the decade.

What the January 2026 BIS Rule Changes

The BIS final rule, effective January 15, 2026, revises the export license review policy for advanced computing semiconductors. Under the new framework, chips with a TPP below 21,000 and DRAM bandwidth under 6,500 GB/s—such as the NVIDIA H200 and AMD MI325X—are now eligible for case-by-case review instead of automatic denial. However, strict conditions apply: a 25% tariff, a 50% volume cap relative to US domestic shipments, mandatory third-party testing in the United States, and know-your-customer (KYC) compliance. Reexport licenses remain under a presumption of denial, signaling that Washington intends to keep a tight leash on the global flow of advanced AI hardware. The rule also closes a previous loophole by requiring that chips be tested in the US before shipment, preventing circumvention through third countries.

According to the Morgan Lewis law firm's analysis, the new policy represents the most significant tightening of AI chip export controls to date. The TPP metric—which measures a chip's theoretical peak performance—is now the central yardstick for determining license eligibility. This shift from blanket denial to conditional case-by-case review is designed to give US regulators more granular control while still allowing some sales under strict oversight. The BIS export control framework now explicitly targets chips that fall just below the most advanced thresholds, closing a gap that Chinese firms had exploited.

China's Retaliation: Rare Earth Export Controls

Beijing's response has been swift and strategic. In early 2026, China implemented its strictest-ever export controls on rare earths and permanent magnets, applying the foreign direct product rule (FDPR) for the first time. Foreign companies must now obtain Chinese government approval to export magnets containing even trace amounts of Chinese-origin rare earth materials or using Chinese technology. Starting December 1, 2025, firms affiliated with foreign militaries face automatic denial of export licenses, and military-use requests are rejected outright. Given that China controls roughly 90% of global rare earth processing, 70% of mining, and 93% of magnet manufacturing, these restrictions pose an existential threat to Western defense supply chains—including the F-35 fighter jet, submarines, and missile systems.

The European Parliament's 2025 briefing notes that over 80% of European companies depend on Chinese supply chains for critical minerals essential to defense, electric vehicles, and renewable energy. Export controls introduced in 2025–2026 have already triggered price spikes of up to sixfold, with licensing approval rates for European firms falling below 25%. The China rare earth export restrictions are not about scarcity but about control: Beijing uses temporary, reversible restrictions to maintain pricing power and extract strategic concessions while discouraging long-term Western investment in alternative supply chains.

Parallel AI Ecosystems and the Fragmentation of Global Innovation

The combined effect of US export controls and Chinese countermeasures is the emergence of two distinct technological spheres. In the United States, companies like NVIDIA and AMD continue to push the frontier of AI hardware, but their access to the Chinese market—historically a major revenue source—is now severely curtailed. The Information Technology and Innovation Foundation (ITIF) estimates that full decoupling could cost US chipmakers $77 billion in lost sales initially, with competitors in South Korea ($21B), the EU ($15B), Taiwan ($14B), and Japan ($12B) capturing much of that market share. US industry R&D investments could decline by about 24% ($14 billion), and the sector could lose over 80,000 direct jobs and nearly 500,000 downstream jobs.

China, meanwhile, is accelerating its domestic semiconductor production. Companies like SMIC and Huawei are ramping up indigenous AI chip architectures, supported by massive state investment. The Chinese government has also imposed a domestic compute mandate, requiring that a growing share of AI computing be performed on domestic hardware. This push for self-sufficiency is creating a parallel ecosystem that, while currently less advanced, could eventually rival US-led innovation. The semiconductor decoupling impact is already visible in the diverging supply chains for advanced packaging, chip design tools, and manufacturing equipment.

Allies Forced to Pick Sides

The US-China chip war is no longer a bilateral dispute—it is reshaping alliances across Asia and Europe. Japan, the Netherlands, and South Korea, all key players in the semiconductor supply chain, are under intense pressure from Washington to align with its export control regime. Japan has already designated 23 items for export controls, and the 2026 MATCH Act in the US seeks to further coordinate restrictions among allies. However, these countries face a dilemma: complying with US demands risks losing access to the Chinese market, which remains a major consumer of chips and chipmaking equipment.

South Korea's Samsung and SK hynix, for example, received annual licenses replacing expired Validated End User (VEU) status in early 2026, but the annual renewal introduces uncertainty. The Netherlands' ASML, the world's sole supplier of extreme ultraviolet (EUV) lithography machines, is caught in the crossfire, with the Dutch government balancing US pressure against economic ties with China. The Japan Netherlands chip export controls coordination has become a central pillar of US strategy, but cracks are appearing as allied firms lobby for more lenient rules.

Expert Perspectives

This is the most consequential rewrite of AI chip export policy since the original October 2022 rules, says Daniel Takahashi, a technology policy analyst. The shift to TPP-based thresholds and case-by-case review gives BIS surgical precision, but it also creates enormous compliance burdens for companies. Meanwhile, China's rare earth countermeasures show that Beijing is willing to play hardball in areas where it holds a monopoly.

The ITIF report concludes that policymakers should keep semiconductor export controls to a minimum to avoid harming US competitiveness. But the geopolitical momentum appears to be moving in the opposite direction, with both sides digging in for a prolonged confrontation.

Frequently Asked Questions

What is the TPP threshold in the 2026 BIS rule?

The new BIS rule sets a total processing power (TPP) threshold of 21,000 for case-by-case review eligibility. Chips below this threshold, such as the NVIDIA H200 and AMD MI325X, can be licensed under strict conditions including a 25% tariff and 50% volume cap.

How is China retaliating against US chip export controls?

China has imposed its strictest-ever export controls on rare earths and permanent magnets, applying the foreign direct product rule (FDPR) for the first time. Foreign firms must obtain Chinese approval to export magnets containing Chinese-origin materials, and military-related requests are automatically denied.

What impact will decoupling have on US chipmakers?

According to ITIF, full decoupling could cost US chipmakers $77 billion in lost sales, reduce R&D investment by 24%, and eliminate over 80,000 direct jobs. Competitors in South Korea, the EU, Taiwan, and Japan are expected to capture much of the lost market share.

How are Japan, the Netherlands, and South Korea affected?

These allies are under pressure to align with US export controls while balancing their own economic interests. Japan has designated 23 items for control, the Netherlands faces pressure over ASML's lithography exports, and South Korea's chipmakers now operate under annual license renewals.

Will the US and China develop separate AI ecosystems?

Yes, the trend points toward parallel AI ecosystems. The US leads in cutting-edge hardware, while China is investing heavily in domestic alternatives and mandating the use of indigenous chips for AI computing. This fragmentation could slow global innovation and increase costs for both sides.

Conclusion: A Defining Strategic Inflection Point

The January 2026 BIS rule and China's rare earth export restrictions mark a defining strategic inflection point for global technology and security policy. The era of globalized semiconductor supply chains is giving way to a bifurcated landscape where technological sovereignty is paramount. For businesses and policymakers, the window to adapt is narrowing. The next 12 to 18 months will be critical in determining whether the world moves toward managed coexistence or full decoupling. The US China technology competition is no longer a future risk—it is the present reality.

Sources

• Morgan Lewis: BIS Revises Export Review Policy for Advanced AI Chips (Jan 2026)
• ECCN Finder: AI Chip Export Controls in 2026
• Rare Earth Exchanges: China's 2026 Export Controls Redraw Global Supply Chain
• CSIS: China's New Rare Earth and Magnet Restrictions Threaten US Defense
• ITIF: Decoupling Risks – Semiconductor Export Controls Harm US Chipmakers (Nov 2025)
• European Parliament: China's Rare Earth Export Restrictions (2025)