The controversy surrounding NVIDIA CEO Jensen Huang could affect the company's bargaining power with the U.S. government and Congress. The area to watch most closely is not lobbying to resume sales of AI chips to China, but negotiations over a congressional bill that would require advanced AI chips to include location-verification features. On November 8, Huang admitted to an extramarital affair and apologized in an email to employees. NVIDIA's board also announced an independent investigation following reports that he had been targeted by a foreign intelligence operation.
According to NVIDIA's quarterly report (Form 10-Q), customers headquartered in China (including Hong Kong) accounted for 8.2% of total revenue in the May–July 2026 quarter. The company also disclosed that it has effectively been shut out of China's data center market, meaning that U.S. export approval alone is not enough to resume sales. Meanwhile, two of the three U.S. House committee chairs who asked Huang for a briefing are also co-sponsors of the "Chip Security Act," which would require location verification for advanced AI chips. Given this, the political impact on NVIDIA may show up less in whether it can sell to China and more in negotiations over the specific conditions of export controls.
Huang's Apology, the Independent Investigation, and the House Request for a Briefing
On November 8, NVIDIA posted the full text of Huang's apology email to employees on its official website. Huang admitted to infidelity toward his wife, apologized to his family, employees, the board, and shareholders, and said he takes full responsibility for his actions. Regarding the national security concerns raised in media reports, he said he would cooperate fully with the independent investigation launched by the board. The email does not name the person with whom he had the affair.
According to the board's statement released the same day, the investigation will be overseen by a committee of independent directors and supported by outside independent legal counsel. The board said it currently has no reason to believe that U.S. national security or NVIDIA's confidential information has been compromised, and it expressed confidence in Huang and the management team. However, the investigation is only getting started, and no final conclusion has been reached on the national security concerns.
The allegations about intelligence activity are, for now, based on media reports. On October 29, The New York Times reported that U.S. intelligence agencies had concluded that an executive at the data center operator Megaspeed, who had been in contact with Huang, had been providing information to China. According to the paper, when President Trump was first briefed in 2025, officials did not believe Huang was aware of this. There is also no information that either party has been indicted, and on October 30, Chinese Foreign Ministry spokesperson Guo Jiakun rejected the report, calling it groundless speculation that slanders China. These allegations cannot be treated as established fact.
Congress, meanwhile, has already asked for an explanation. According to an announcement from the House Foreign Affairs Committee, Foreign Affairs Committee Chairman Brian Mast, Select Committee on China Chairman John Moolenaar, and Intelligence Subcommittee Chairman Rick Crawford sent a letter dated November 4 asking Huang to attend a joint closed-door briefing on November 18. The letter asks him to explain his contacts with a person described in media reports as a Chinese intelligence asset, as well as NVIDIA's measures to protect sensitive information. They asked Huang to respond by November 11 on whether he will attend or to propose alternative dates, but as of now his attendance on November 18 is not confirmed.
China-Headquartered Customers Account for 8.2% of Revenue; Data Center Market Effectively Closed
According to NVIDIA's 10-Q for the May–July 2026 quarter, filed with the U.S. Securities and Exchange Commission (SEC), revenue from customers headquartered in China (including Hong Kong) was $7.88 billion, about 8.2% of total revenue of $96.221 billion. Converted at ¥157.75 to the dollar, the Tokyo market rate on November 6, $7.88 billion is roughly ¥1.2431 trillion. For the first half cumulatively, the share of revenue from China-headquartered customers was 7.0%, down sharply from 15.0% in the same period a year earlier. However, this is regional revenue based on where direct customers are headquartered; it does not indicate actual shipping destinations or the locations of end customers.
Therefore, the 8.2% figure cannot be used to gauge end demand within China or the scale of exports routed through third countries. NVIDIA itself states in the 10-Q that the headquarters location of a direct customer can differ from the location of the end customer or the product's shipping destination. In addition, against data center revenue of $89.023 billion in the quarter, shipments of Hopper products to China were less than 1%. This figure uses data center revenue as its denominator, so it covers a different base from the 8.2% share of total revenue from China-headquartered customers. It is also not disclosed how much of that 8.2% corresponds to products for data centers inside China.
The history of restrictions on AI chip sales to China shows why U.S. government permission alone is not enough to restart the business. Since February 2026, the U.S. government has issued licenses allowing limited shipments of H200 chips to certain Chinese companies. However, the approved products must undergo inspection in the United States before shipment, and the import into the U.S. for that purpose is subject to a 25% tariff. Furthermore, because the Chinese government restricted purchases, NVIDIA was unable to sell the full approved volume, and in the first half of fiscal 2026 it recorded $400 million in charges related to excess H200 inventory and purchase obligations.
In the 10-Q, NVIDIA explains that it cannot broadly supply competitive products approved by both the U.S. and Chinese governments to China's data center market. As a result, it acknowledged that, as of the end of the May–July 2026 quarter, it was effectively shut out of China's data center computing market. Even if the U.S. government permits H200 exports, it is difficult to restart sales in earnest as long as Chinese government purchase restrictions continue. Even if Huang's political influence changes, this business environment will not improve immediately.
That said, this does not mean NVIDIA places little importance on returning to the Chinese market. The company can ship GPUs that are not subject to export controls, such as those for gaming and workstations, to China. But it warns that if it cannot return with competitive data center products, lost business opportunities and the growth of competitors could materially harm its results. The motivation to return to China remains strong, but doing so requires approval from both the U.S. and China. For this reason, in considering how this controversy affects NVIDIA's political influence, it seems more appropriate to focus on negotiations over export conditions that can be decided within the U.S., rather than on the resumption of China sales itself.
Two of the Chairs Who Requested a Briefing Co-Sponsored the Bill Mandating AI Chip Location Verification
The Chip Security Act (H.R.3447), being advanced in Congress, was introduced on May 15, 2025. As introduced, the text would require the U.S. Secretary of Commerce, within 180 days of enactment, to mandate that covered advanced AI chips and related products include location-verification features. The covered items are products classified under U.S. Export Control Classification Numbers (ECCN) such as 3A090, 3A001.z, 4A090, and 4A003.z, and the rules would apply not only to exports but also to re-exports and in-country transfers within destination countries.
A distinguishing feature of the bill is that it is not limited to specific manufacturers or to products bound for China. The covered products are defined by ECCN, and the text as introduced contains no provision blanket-exempting allied countries. Therefore, if it were enacted in this form, products imported by data center operators in Japan would also need location-verification features if they fall under covered classifications such as 3A090 or 4A090. By also covering re-exports and in-country transfers after export from the U.S., the mechanism aims to prevent products from straying outside approved channels.
The bill also requires businesses that have received export licenses or similar approvals to promptly report to authorities if they obtain credible information that a product's location or users differ from what was stated in the application, or that someone has attempted to tamper with or disable the location-verification feature. In other words, beyond giving chips location-verification capability, the bill sets up a system to monitor and report on whether the approved conditions of use are being followed. Note, however, that this describes the text as introduced in May 2025; the final regulatory content following subsequent deliberation and amendment has not been determined.
Co-sponsors of the bill include Republicans John Moolenaar and Rick Crawford, along with Democrats Raja Krishnamoorthi, Ted Lieu, Josh Gottheimer, and Bill Foster. Moolenaar and Crawford are two of the three chairs who requested the briefing from Huang. In addition, the House Foreign Affairs Committee, chaired by the third, Brian Mast, passed the bill in committee on March 26, 2026 and decided to advance it toward the full House. It is significant that the lawmakers who requested Huang's briefing are simultaneously advocates of tighter export controls on AI chips.
In the risk factors of its 10-Q, NVIDIA notes that the U.S. government could in the future require chip-tracking or performance-limiting features as a condition of export licenses. For example, it could be required to build a mechanism that disables or restricts GPU functions when an unauthorized system configuration, use, or location is detected. The company warns that mandating such features could create new vulnerabilities and harm demand for its products and its business. However, mandating location verification is not the same as mandating the ability to remotely disable GPUs, so the bill's requirements need to be distinguished from the future regulations NVIDIA is concerned about.
Separately, in an official blog post on December 10, 2025, NVIDIA described software that lets data center operators centrally manage GPU operating status and configuration. It explained that the system is optional for customers to adopt, that its telemetry data is read-only, and that it has no ability to remotely change GPU settings or operation. NVIDIA also stressed that its GPUs have no hardware-based tracking features, kill switches, or backdoors.
Given this, the future points of contention are not only whether location verification can be technically implemented. What matters are the conditions: how far the mandated functionality should go, how covered products and export destinations are defined, and who carries out actual monitoring and reporting. The congressional side that sets those conditions includes the very chairs who have now asked Huang for an explanation. If this controversy weakens NVIDIA's bargaining power, the effect is most likely to appear in negotiations over these regulatory conditions.
Tighter Regulation Was Already Under Way; the Final Defense Authorization Bill Is the Focus
That said, there is an important counterargument to the view that this controversy will advance a location-verification mandate. The Chip Security Act was introduced in May 2025 and passed the House Foreign Affairs Committee in March 2026, before the reports about Huang emerged. Moreover, in the committee's announcement, Rep. Bill Huizenga explained that the bill supports the Trump administration's AI Action Plan and incorporates the administration's technical assistance and industry input. In other words, the movement toward a location-verification mandate has proceeded independently of this controversy, and it cannot be asserted that the allegations around Huang will strengthen regulation.
There are also political constraints. As NPR reported, President Trump holds veto power over legislation, so a bill passed by Congress would in principle need the president's support to become law. Ahead of the midterms, Trump had signaled that he would actively use the veto if Democrats took control of Congress. However, that remark concerned Democratic policies and did not name the Chip Security Act. Trump has also made remarks defending Huang, and NVIDIA's board says it currently has no reason to believe confidential information was compromised. For these reasons, it is hard to imagine this controversy immediately costing NVIDIA its political influence.
On the other hand, the balance of power in Congress has changed significantly. According to AP's tally, Democrats won majorities in both chambers in the November 3 U.S. midterm elections, and when the new Congress convenes in January 2027, control will be divided between the Trump administration and the congressional majority. Democrats have signaled they will increase oversight of the administration and large corporations, and the NVIDIA matter could be pursued further in Congress. Even so, even if the new Congress is eager to tighten regulation, the constraint of the presidential veto does not disappear.
Against this political backdrop, what matters in the short term is the National Defense Authorization Act (NDAA) for fiscal 2027, expected to be deliberated through the end of the year. The NDAA is an annual law setting the framework for U.S. defense policy and the defense budget, and it often includes related security policy provisions. On November 6, Semafor reported that export control provisions further restricting sales of AI chips to China could be added to defense legislation considered during the lame-duck session between the midterms and the start of the new Congress. The NDAA is a must-pass bill, and it may offer an easier path to enacting related regulations than considering them on their own. However, what Semafor reported concerned restrictions on chip sales to China; it has not been confirmed that a location-verification mandate will be included in the final bill.
The upcoming schedule is also important. Following the November 4 congressional request for a briefing, NVIDIA announced Huang's apology and the start of the board's independent investigation on the 8th. The deadline for responding to Congress is the 11th, NVIDIA's fiscal 2027 third-quarter results are due on the 17th, and the closed-door briefing requested by Congress is set for the 18th. These are concentrated within about two weeks, with the start of the new Congress in January 2027 beyond them. However, November 18 is the date Congress wants for the briefing, and whether Huang will attend has not yet been decided.
To test the view that this controversy weakens NVIDIA's bargaining power, it will be necessary to watch upcoming earnings and congressional developments. First, the fiscal 2027 third-quarter results (quarter ended October 25), to be announced on November 17 (U.S. time), and the 10-Q filed afterward will serve as reference points. If shipments of Hopper products and H200 to China rise significantly from the previous quarter's "less than 1% of data center revenue," or if the share of revenue from China-headquartered customers substantially exceeds 8.2%, the importance of China sales will need to be reassessed. However, because the revenue share based on customer headquarters cannot capture demand from data centers inside China, it must be checked alongside actual shipment data.
On congressional developments, the key question is whether the Chip Security Act is put to a vote on the House floor before the new Congress convenes in January 2027. If the session ends without a vote, it will be difficult for the bill to become law as a standalone measure in the current Congress, and the short-term impact of this controversy would likely be limited. More important is whether the final NDAA text includes provisions on location verification or chip tracking. If such provisions are not included, the view that reduced NVIDIA bargaining power will directly lead to tighter regulation weakens.
What is at stake in this controversy is not only whether NVIDIA can sell AI chips to China. As introduced, the location-verification mandate being considered in Congress could affect overseas markets, including Japan, so it also bears on the operating conditions of data center operators and AI infrastructure worldwide. A causal link between the allegations around Huang and tighter regulation cannot be confirmed at this point, and the results of the independent investigation are not yet in. The focus going forward is how well NVIDIA can maintain its position in negotiations with Congress over export conditions, and whether the final NDAA text includes location-verification or chip-tracking requirements.