The Global Fever

Policy

US Export Controls on Consumer Chips Widen as Bessent Says Nvidia Curbs Are Enough

A Commerce rule in force since November 19 now reaches chips used in consumer devices, while Treasury Secretary Bessent says Nvidia's older chips need no further limits.

Wasif Mumtaz

Founder and Editor · November 21, 2026 · 7 min read

Close-up of a circuit board with a central microchip and small surface-mounted components under purple light
Photo: Magnific

US export controls on consumer chips are now part of the story, not just rules for advanced AI processors. A Commerce Department rule that took effect on November 19 reaches chips used in everyday mobile and consumer devices, according to Nikkei Asia, and TSMC has paused some shipments while customers seek licences. A day later, Treasury Secretary Scott Bessent told CNBC that no further restrictions on Nvidia's older chips are needed.

The two developments point in different directions. One widens the reach of Washington's controls across the electronics supply chain. The other says the government has already done enough on AI chips. We covered the November 16 rule in our report on US chip export restrictions and Nvidia's H200 halt and the reaction from Alibaba. This piece covers what is new since then.

What the new US export controls on consumer chips change

The Commerce Department's Bureau of Industry and Security issued an interim final rule on November 16, and it took effect on November 19 after Federal Register publication. The headline feature was a crackdown on remote access to advanced AI chips. TechCrunch reported that it requires licences for AI chips going to data centres in Malaysia, Singapore, Indonesia, Thailand and Brazil, and that it rescinds the AI Diffusion Rule issued in January 2025.

What surfaced this week is how far the rule reaches. Nikkei Asia reported on November 20 that the controls extend beyond advanced AI chips to chips used in everyday mobile and consumer applications. It described the impact as far broader than first expected and said developers outside China are affected, not only Chinese firms. The report draws on unnamed industry executives.

A Reuters report, republished by the Taipei Times and citing two people familiar with the matter, explains the mechanism. TSMC used to be able to verify a chip's computing power to show it fell outside US export controls. The new rule removed that route, so advanced chips for affected Chinese customers are now presumed to be covered. Reuters said the Financial Times first reported the shipment suspensions.

How the rule reaches smartphone chips and the supply chain

According to Reuters, TSMC is halting some previously permitted orders at 16-nanometre processes and below unless customers obtain US export licences. Chinese chip designers will generally be limited to 28-nanometre and 40-nanometre processes, and only if their chips do not use high-bandwidth memory or need advanced packaging. Those are the more mature nodes that sit well behind the leading edge.

TSMC is also widening its customer checks. Reuters said the company is scrutinising operating locations, parent companies, major shareholders, packaging providers and shipping routes. The checks extend to orders placed through overseas subsidiaries or intermediaries in Taiwan, the United States, Europe, Japan and Southeast Asia. The US list of approved designers includes Nvidia, AMD, Apple, MediaTek and Realtek, but no mainland Chinese firms.

The consumer angle comes through most clearly in a smaller outlet. The Seoul Economic Daily reported on November 21 that the licensing requirement now extends to consumer smartphone chips for companies such as Xiaomi, Oppo and Vivo, and that TSMC's Chinese smartphone supply chain is effectively under full review. It said licences for consumer chips are expected to be approved eventually, but that production disruptions from procedural delays are unavoidable.

The same outlet reported that Samsung Electronics has halted shipments of advanced chips to Baidu's Kunlun AI unit and to ByteDance, and it put Samsung's China revenue exposure at about 20%. We could not confirm those company-level details in a second source, so treat them as reported rather than official. Reuters also said it could not independently verify how some Chinese AI chips were made at TSMC, and the sources we reviewed do not include a published list of affected orders.

For electronics makers, the practical issue is timing. A licence that is likely to be granted still takes time to obtain, and chip orders are scheduled months ahead. A pause at the foundry can delay a phone or device build even if the final answer is yes. That is the risk the Seoul Economic Daily flagged when it said delays are unavoidable.

What Bessent said about Nvidia chip exports to China

The Associated Press reported that Bessent spoke on Friday, November 20, in a CNBC interview after returning from an Asia trip with President Donald Trump. He said no further restrictions on sales of older Nvidia chips to China are needed, now that the administration has blocked China's remote access to the latest semiconductors. In his words, "The proper chip restrictions are in place," and "the remote access has been curtailed."

The scope matters. AP said the comments concern Nvidia's H200, one of its more powerful products but not the most advanced, along with AMD's roughly comparable MI325X. So the statement is narrower than the headline suggests. It addresses older chips that can be sold under licence, and it does not loosen the new rule on remote access or foundry orders.

AP also cited a House Foreign Affairs Committee report showing that the Bureau of Industry and Security had approved 879,868 chips across 21 licence applications for companies with Chinese ties as of May 11, 2026. Nvidia and AMD had asked to export more than 3.4 million. AP noted that an approval does not necessarily mean a completed sale. The approved amounts included 75,000 for Alibaba and 150,000 for Tencent and its subsidiaries.

Nvidia told AP that the limited shipments it made under the H200 program were legal and authorized, and that the review and licensing process lasted more than half a year. The company said it stopped shipping H200 chips to China earlier that week. AMD said it complies with US export control laws.

The interview also exposed friction inside the administration. AP reported that Bessent said Treasury wanted export controls placed on Megaspeed, a Singapore-based company the US is investigating over the possible diversion of restricted Nvidia chips to China, but that Commerce was in charge of the matter. AP described him as frustrated by the delay. President Trump said he has not lost confidence in Commerce Secretary Howard Lutnick.

What is actually new this week and what to watch

Three things are new. The rule's reach into consumer and mobile chips came into focus on November 20 and 21. TSMC, and reportedly Samsung, began pausing or tightening China-linked orders. And Bessent publicly signalled that Washington does not plan to go further on Nvidia's older chips. The November 16 rule itself, Nvidia's H200 halt and Alibaba's warning were already in the news.

Several questions remain open. The first is how quickly licences for consumer chips are processed, since that decides how much production slips. The second is whether Commerce clarifies the consumer scope, because the rule is an interim final rule and its details can still change. TechCrunch reported that existing data centre operators serving covered customers have 90 days from the effective date to get licences, so a first compliance deadline arrives in the new year.

China's response is the third. Digitimes reported that Beijing vowed to retaliate against the expanded controls, though the details of any countermeasures were not available to us. This article reports statements and press coverage and is not investment advice.

Frequently asked questions

Do the new US export controls cover consumer electronics chips?

Nikkei Asia reports that the controls go beyond advanced AI chips to chips used in everyday mobile and consumer applications. The Seoul Economic Daily says smartphone chips for Xiaomi, Oppo and Vivo now face licensing. Neither report says consumer chips are banned, and licences are expected to be granted eventually.

When did the rule take effect?

The Bureau of Industry and Security issued the interim final rule on November 16, 2026. It took effect on November 19, when it was published in the Federal Register.

Did Bessent say Nvidia can sell any chip to China?

No. AP reports he said no further restrictions are needed on older Nvidia chips, such as the H200, after remote access was curtailed. He did not say the latest chips can be sold, and the new licensing rules remain in force.

Which companies have paused shipments?

Reuters, following the Financial Times, reports that TSMC suspended some previously permitted orders at 16-nanometre and below unless licences are obtained. The Seoul Economic Daily reports that Samsung halted shipments to Baidu's Kunlun unit and ByteDance. The Samsung detail comes from a single smaller outlet.

Has China responded?

Digitimes reported that China vowed to retaliate against the expanded controls. We did not find details of specific countermeasures in the sources we could read, so the form any response takes is not yet clear.

ShareXLinkedInWhatsApp

Sources

Related stories

Policy

Trump Threatens Tax on AI Data Centers Abroad as Senate Weighs AI Committee

Trump says US firms that build AI data centers abroad will face "a tremendous tax" but gave no details. Sen. Hickenlooper separately proposed a Senate select committee on AI.

Policy

Fed Rate Hike Case Grows as Paulson and Goolsbee Voice Inflation Worry

Philadelphia Fed's Anna Paulson backs more gradual increases, while Chicago Fed's Austan Goolsbee has left a larger December move open, citing AI-driven demand.

Policy

US Chip Export Restrictions Tighten as Nvidia Halts China H200 Shipments

A new Commerce rule closes the remote-access loophole for advanced AI chips, and Nvidia says it will stop shipping H200s to China. Here is what changes for chipmakers and cloud operators.