·The Hindu

How U.S. curbs flow of latest chips to China

FlashcardsQuick reviseKey facts. One by one.Start →
In this note
  1. At a Glance
  2. Why in the News
  3. Background & Evolution
  4. Core Static Facts
  5. Multi-Dimensional Analysis
  6. Recent Developments (last 12–18 months)
  7. Prelims Hooks
  8. Why Chips Still Reach China Despite the Ban
  9. Washington Tightens and Loosens at the Same Time
  10. The Strongest Case Against the Curbs, and an Honest Answer
  11. What the Rule Means for India's Chip and AI Plans
  12. Anchors for Answers
  13. Mains Relevance
  14. Related Topics to Study Next
  15. NCERT Concepts to Revise
  16. Common Errors / Trap Areas

1. At a Glance

2. Why in the News

  • On 16 November 2026, the U.S. Commerce Department issued an Interim Final Rule (IFR). It requires companies to hold licences to supply advanced American AI chips to data centre operators outside the U.S. that have substantial computing power and serve businesses headquartered in two dozen countries, including China, Iran and Russia [1].
  • The rule also stops American firms from freely shipping high-end AI chips to a set of five countries, including Indonesia, Malaysia and Singapore, without explicit U.S. government approval [1].
  • China condemned the move. It called it "yet another case of the U.S. overstretching the concept of national security, and abusing export control measures" to interfere in China's trade and tech cooperation with other countries and to go after China's high-tech industry [1].
  • Timing: the rule came days before a meeting between U.S. President Donald Trump and Chinese President Xi Jinping in Shenzhen, China [1].

3. Background & Evolution

  • Origin: U.S. export controls on chips to China began during the first Trump term [1].
  • Since then, semiconductor technology and business have changed significantly, especially with the AI boom [1].
  • How the controls have widened:
  • Phase 1: controls on specific Chinese firms and chipmaking technology. [Verify] Entity List actions, e.g. Huawei, 2019.
  • Phase 2: broad performance-based curbs on advanced computing chips and chipmaking equipment. [Verify] The October 2022 rules, tightened in 2023 and 2024.
  • Phase 3: a shift from controlling chips to controlling compute access, i.e. cloud services provided from data centres in other countries. This is the November 2026 IFR [1].
  • Phase 4: licensing for transit and transhipment hubs in Southeast Asia (Indonesia, Malaysia, Singapore) to stop chips being diverted to China [1].

  • India's parallel track: the Union Cabinet approved ISM 1.0 in December 2021 to build a domestic chip ecosystem [2].

4. Core Static Facts

Item Fact
U.S. issuing authority U.S. Department of Commerce [1]. [Verify] Its Bureau of Industry and Security (BIS) administers the Export Administration Regulations (EAR).
Instrument Interim Final Rule (IFR): in force on issue, with comments invited afterwards [1]
Date of latest rule 16 November 2026 [1]
Who is regulated Data centre operators outside the U.S. equipped with advanced American AI chips [1]
Restricted end-users Businesses headquartered in about 24 countries, including China, Iran, Russia [1]
Transit-country licensing 5 countries, including Indonesia, Malaysia, Singapore [1]
China's objection "Overstretching" national security; "abusing export control measures" [1]
India's response track India Semiconductor Mission (ISM); ISM 1.0 approved by Cabinet in Dec 2021 [2]
ISM outlay ₹76,000 crore, for fabs, display fabs and related investments [3]
ISM 2.0 Announced in Union Budget 2026–27; covered in a PIB backgrounder dated 7 Feb 2026 [2]
NITI Aayog input Report Future of India Semiconductor Industry, May 2026 [4]

Key terms:

  • Export control: a government licensing requirement for exporting strategic goods or technology.
  • Interim Final Rule: a rule that takes effect immediately, before the public comment process.
  • Compute / computing power: the processing capacity that AI training depends on.
  • Transhipment / diversion: re-routing controlled goods through third countries.
  • Extraterritoriality: applying a country's law to entities outside its borders, in this case foreign data centres [1].

5. Multi-Dimensional Analysis

Geopolitical / Strategic

  • Advanced chips are now treated as a strategic commodity. The article calls them "the most-sought-after commodity" underpinning AI [1].
  • Southeast Asian hubs (Singapore, Malaysia, Indonesia) are pulled into the U.S.–China contest, which narrows their room to stay neutral [1].
  • The rule was timed just before the Trump–Xi Shenzhen summit, which suggests it is also meant as negotiating leverage [1].
  • For India, the rivalry is a chance to present itself as a "trusted geography" in chip supply chains. This is the strategic logic behind ISM [2][3].

Economic

  • Licensing adds compliance costs for cloud and data-centre operators around the world and could split the global AI-compute market [1].
  • China has framed the curbs as interference in "normal trade and tech cooperation," which feeds wider de-globalisation trends [1].
  • India's ₹76,000-crore ISM aims to attract fab and display investment as supply chains shift [3].

Legal / Governance

  • The rule reaches beyond U.S. borders: foreign data centres are regulated because of where their chips came from [1].
  • China's criticism raises the long-running WTO debate over the national-security exception versus free trade. [Verify] GATT Article XXI.
  • Using an IFR, with no prior notice and comment, gives the executive speed but less consultation [1].

Scientific / Technological

  • Controls have moved from hardware to cloud and remote access, closing the loophole of renting compute abroad [1].
  • Semiconductor technology and business have changed a great deal since the first controls, which is why controls keep being revised [1].
  • India's ISM 2.0 focuses on the next layer of capability beyond assembly [2][4].

Ethical

  • The rule pits national security against open science and access to technology. Restrictions on compute may widen the global AI divide [1].

6. Recent Developments (last 12–18 months)

  • Aug 2025: a PIB document restated ISM's ₹76,000-crore outlay for fabs, display and related investments [3].
  • Feb 2026: ISM 2.0 was announced in Union Budget 2026–27, with a PIB backgrounder dated 7 Feb 2026 [2].
  • May 2026: NITI Aayog released Future of India Semiconductor Industry [4].
  • 16 Nov 2026: the U.S. Commerce Department issued the IFR on licensing data centre and cloud services, and licensing for five transit countries [1].
  • Nov 2026: China condemned the rule; the Trump–Xi meeting in Shenzhen followed days later [1].

7. Prelims Hooks

  • The U.S. rule of 16 November 2026 is an Interim Final Rule, issued by the U.S. Commerce Department [1].
  • It regulates data centre operators outside the U.S. that use advanced American AI chips [1].
  • It restricts cloud services to businesses headquartered in about two dozen countries, including China, Iran, Russia [1].
  • It adds a licence requirement for chip shipments to five countries, including Indonesia, Malaysia, Singapore [1].
  • U.S. chip export controls on China date back to the first Trump term [1].
  • The Trump–Xi meeting after the rule was held in Shenzhen, China [1].
  • ISM 1.0 was approved by the Union Cabinet in December 2021 [2].
  • ISM's financial outlay is ₹76,000 crore [3].
  • ISM 2.0 was announced in Union Budget 2026–27 [2].
  • NITI Aayog's Future of India Semiconductor Industry report came out in May 2026 [4].
  • ISM's nodal ministry is MeitY (Ministry of Electronics and IT), not the Ministry of Commerce [2].

8. Why Chips Still Reach China Despite the Ban

  • Banned chips are being smuggled through other countries
  • In December 2025, U.S. authorities detained two people for allegedly smuggling Nvidia AI chips worth $160 million to China [8].
  • In 2026, prosecutors charged a co-founder of Super Micro, a U.S. server maker. It is the biggest chip smuggling case since the U.S. first restricted Nvidia shipments to China in 2022 [7].
  • So a ban on paper does not stop chips at the border. Chips are sold to a firm in a third country and then quietly moved on to China. This is why the new rule adds licences for transit hubs [1].

  • Renting computing power was a bigger gap than smuggling

  • A Chinese firm does not need to own a chip to use it. It can rent computing power from a data centre in another country through the cloud.
  • One think-tank estimate says this remote access could make up about a quarter of China's total AI computing capacity by the end of 2026 [6].
  • Until this rule, the U.S. did not need a licence to ship Nvidia AI chips to countries like Malaysia, Singapore and Thailand [6]. Those countries were the easiest route.

  • The U.S. still cannot see where a chip goes after it is sold

  • Export licences control the first sale. They do not track the chip afterwards.
  • U.S. lawmakers have proposed a Chip Security Act. It would require companies to track the location of advanced AI processors so that smuggling can be detected [10].
  • Until something like this becomes law, enforcement depends on catching smugglers after the chips have already left.

9. Washington Tightens and Loosens at the Same Time

  • The same government that blocks cloud access has approved large chip sales
  • The Trump administration approved exports of Nvidia's H200 chips to China. This reversed restrictions imposed under President Biden [5].
  • House Foreign Affairs Committee documents show the Commerce Department approved shipments of more than 879,000 advanced AI chips to Chinese companies between December 2025 and May 2026 [5].
  • So the November 2026 rule closes one door (cloud access) while another door (direct sales of slightly older chips) was opened months earlier.

  • This makes the controls look like a bargaining chip, not a fixed security line

  • The rule came days before the Trump–Xi meeting in Shenzhen [1].
  • When controls are loosened and tightened in step with summits, they read as negotiating tools. This supports China's charge that the U.S. is "overstretching" national security [1].
  • U.S. lawmakers are pushing for tougher curbs at the same moment [10]. This shows that Congress and the White House do not agree.

  • Why this matters for third countries like India

  • The Biden-era global framework for AI chip exports was scrapped by the Trump team [6]. It has been replaced by country-by-country rules.
  • A data-centre investor in India or Southeast Asia cannot be sure which rule will apply in two years. That uncertainty is itself a cost of doing business.

10. The Strongest Case Against the Curbs, and an Honest Answer

  • The objection: the curbs push China to build its own chips faster
  • Chinese regulators have told domestic tech companies to stop ordering Nvidia's China-specific chips [9].
  • China has also tightened checks on chip imports to reduce its reliance on U.S. technology [9].
  • So the U.S. may lose its Chinese customers and its leverage at the same time, while China builds rival chipmakers such as Huawei [9].

  • The objection: everyone pays for a divided tech world

  • The IMF estimates that trade fragmentation (the world splitting into rival trading blocs) could cost from 0.2% of global GDP in a mild scenario to 7% in a severe one [11].
  • When technology decoupling is added, losses in some countries reach 8–12% of GDP [11].

  • The honest answer: the objection is partly right, but the cost in chips falls mostly on China

  • An IMF working paper modelled export bans between China and OECD countries. In semiconductors, China's chip quality fell by about 5%, while the OECD saw almost no loss [12].
  • Leading chips are hard to replace, and building new supply is slow and costly [13]. That is why the curbs hurt in the short run.
  • What the critics get right: the curbs buy time, not a permanent lead. China's self-reliance push shows the gap will narrow. The curbs also raise costs for neutral countries that had no part in the dispute.

11. What the Rule Means for India's Chip and AI Plans

  • India is not on the five-country transit list, for now
  • The countries needing licences for chip shipments are Singapore, Thailand, Malaysia, Indonesia and Brazil [6].
  • This gives India a short-term edge as a place to build AI data centres. Its Southeast Asian rivals now face extra paperwork [6].
  • But Indian data centres using American chips would still need licences to serve firms headquartered in the two dozen "countries of concern" [1].

  • One modern chip factory costs almost the whole ISM budget

  • A leading-edge fab (chip factory) needs an upfront investment of US$10–20 billion [13].
  • ISM's total outlay is ₹76,000 crore [3]. At today's exchange rates that is under US$10 billion, and it must also cover display fabs and other projects.
  • So India cannot match the U.S. or China at the most advanced chips with ISM money alone. The realistic aim is to become a trusted partner in parts of the supply chain.

  • MeitY and DGFT should show the U.S. that chips in India stay in India

  • The U.S. moved Singapore and Malaysia onto the licence list because chips were leaking through them [6][7].
  • India can avoid the same fate by tracking where imported AI chips end up. This could be done under its SCOMET list (India's own list of dual-use goods that need export permission). The U.S. Chip Security Act proposal uses the same idea of location tracking [10].
  • A clean record is what turns India's "trusted geography" claim into something the U.S. can verify.

  • The IndiaAI Mission depends on U.S. licensing decisions

  • India's public GPU pool depends on advanced American chips. U.S. policy has changed direction several times since 2025 [5][6].
  • This is the strongest argument for ISM 2.0's push beyond assembly [2]. Self-reliance in chips is also protection against shifting U.S. rules.

12. Anchors for Answers

  • Data: More than 879,000 advanced AI chips approved for export to Chinese firms, Dec 2025–May 2026 (House Foreign Affairs Committee documents) [5]
  • Data: Remote cloud access may supply about a quarter of China's AI computing capacity by end-2026 (think-tank estimate) [6]
  • Data: A leading-edge fab costs US$10–20 billion upfront (OECD, 2025) [13]
  • Data: Geo-economic fragmentation could cost 0.2%–7% of global GDP; 8–12% in some countries once tech decoupling is added (IMF, 2023) [11]
  • Report/Committee: IMF Staff Discussion Note, Geo-Economic Fragmentation and the Future of Multilateralism (2023) [11]; OECD, Economic Security in a Changing World (2025) [13]; NITI Aayog, Future of India Semiconductor Industry (2026) [4]
  • Law/Case: Proposed U.S. Chip Security Act, which would require location tracking of advanced AI chips [10]; U.S. prosecution of a Super Micro co-founder for chip smuggling [7]
  • Comparison: China responded to U.S. curbs by telling its firms to stop buying Nvidia's China-specific chips and by tightening import checks, which speeds up its own self-reliance push [9]
  • Scheme: India Semiconductor Mission (₹76,000 crore) and ISM 2.0, India's route to becoming a trusted chip-supply partner [2][3]

13. Mains Relevance

14. Related Topics to Study Next

  • India Semiconductor Mission 1.0 and 2.0: India's direct policy response to the reshaping of the chip supply chain [2][3].
  • Production Linked Incentive (PLI) schemes for electronics: the companion industrial-policy tool to ISM.
  • SCOMET list and India's export-control regime: India's own dual-use technology controls.
  • Wassenaar Arrangement, MTCR, Australia Group: multilateral export-control regimes India belongs to.
  • iCET / TRUST initiative (India–U.S.): bilateral cooperation on critical and emerging technologies.
  • WTO national-security exception (GATT Art. XXI): the legal battleground for China's objections.
  • Critical minerals and China's rare-earth/gallium export curbs: China's counter-leverage in the supply chain.
  • IndiaAI Mission and GPU compute access: India's own need for advanced chips and exposure to U.S. licensing.

15. NCERT Concepts to Revise

  • Export controls (Economy): The U.S. rule is an export control: licensing sensitive chips and cloud access to curb China's AI progress.
  • Weaponised interdependence (Economy): Shows how control over chokepoint technology like advanced chips gives the U.S. leverage over rivals and third countries.
  • Security exception (Economy): China calls the rule an abuse of national security; this WTO clause is the legal battleground.
  • World Trade Organization (History): Explains the trade-rule body before which China's objections to U.S. export curbs would be tested.
  • General Agreement on Tariffs and Trade (History): GATT Article XXI, the original national-security exception, underlies the legal debate over these curbs.
  • Artificial intelligence (Economy): Advanced chips matter because they supply the computing power for AI, the rivalry's core prize.
  • Cloud computing (Economy): The new rule targets remote access to chips via the cloud, not just physical shipments.
  • Data centre (Geography): Foreign data centres using American AI chips are the entities the rule regulates.
  • Supply chain resilience (Economy): Chip curbs and diversion risks push countries to build resilient, diversified semiconductor supply chains.
  • Dual-use goods (Economy): Advanced chips serve civilian and military uses, which is why they fall under export controls.
  • Singapore (Geography): Singapore is a key transhipment hub now needing licences to stop chip diversion to China.
  • India Semiconductor Mission (Geography): India's main policy response to build a domestic chip ecosystem amid this rivalry.
  • Self-reliance (Economy): India's semiconductor push aims at self-reliance, reducing dependence on foreign chip supplies and controls.
  • Semiconductor fabrication plant (Geography): Fabs are the costly chipmaking step that ISM seeks to attract to India.
  • IndiaAI Mission (Economy): India's AI compute pool needs advanced GPUs, exposing it to U.S. licensing rules.

16. Common Errors / Trap Areas

  • Wrong U.S. agency: the rule was issued by the Commerce Department, not the State Department, Treasury or USTR [1].
  • Scope confusion: the November 2026 rule targets cloud and data-centre services abroad, not just direct chip exports to China [1].
  • Wrong country lists: "two dozen countries" (China, Iran, Russia) are restricted end-user headquarters. The "five countries" (Indonesia, Malaysia, Singapore) are shipment destinations needing licences. Do not merge the two lists [1].
  • Wrong origin: the controls began in the first Trump term, not under Biden. Biden-era rules expanded them [1].
  • Wrong ISM year and ministry: ISM 1.0 dates to Dec 2021 under MeitY. ISM 2.0 comes from Budget 2026–27. The outlay is ₹76,000 crore, which is not the same as the PLI total [2][3].

Sources

  1. 1How U.S. curbs flow of latest chips to China, John Xavier, The Hindu (22 Nov 2026, p. 12):thehindu.com · tier 4
  2. 2PIB Backgrounder: India Semiconductor Mission 2.0, A Major Push in Budget 2026 towards Semiconductor Self-Reliance:pib.gov.in · tier 1
  3. 3PIB document (Aug 2025) on the India Semiconductor Mission outlay:static.pib.gov.in · tier 1
  4. 4NITI Aayog, Future of India Semiconductor Industry (May 2026):niti.gov.in · tier 1
  5. 5Trump admin cleared over 879k AI chip sales to Chinese firms: House panelbusiness-standard.com · tier 4
  6. 6US moves to restrict China's remote access to AI chips after Nvidia rowbusiness-standard.com · tier 4
  7. 7US charges Super Micro co-founder with smuggling AI chips to Chinabusiness-standard.com · tier 4
  8. 8US detains two for allegedly smuggling $160 mn Nvidia AI chips to Chinabusiness-standard.com · tier 4
  9. 9China tightens chip import checks to curb reliance on Nvidia, US techbusiness-standard.com · tier 4
  10. 10US lawmakers push for tougher curbs on advanced AI chip exports to Chinabusiness-standard.com · tier 4
  11. 11IMF Staff Discussion Note: Geo-Economic Fragmentation and the Future of Multilateralism (2023)elibrary.imf.org · tier 2
  12. 12IMF Working Paper WP/24/122: The Price of De-Risking: Reshoring, Friend-Shoring, and Quality Downgradingimf.org · tier 2
  13. 13OECD, Economic Security in a Changing World (2025): Special focus: Semiconductor value chainsoecd.org · tier 2

Mains Q&A on this note

NCERT concepts in this note

Also on 22 November

All 22 November articles →