By Stephen Nellis and Max A. Cherney
SAN FRANCISCO () - In the upcoming weeks, a U.S. legislator intends to propose new legislation aimed at tracking the whereabouts of AI chips similar to those produced by Nvidia once they have been distributed for sale.
The effort to keep tabs on the chips, which drew bipartisan support from U.S. lawmakers, aims to address reports of widespread smuggling of Nvidia's chips into China in violation of U.S. export control laws.
NVIDIA’s processors are essential components in developing AI systems like chatbots and image generators, as well as more niche applications that could potentially aid in creating bioweapons. Both President Donald Trump and former president Joe Biden have successively enforced stricter export restrictions on NVIDIA’s chips destined for China.
During Nvidia’s previous financial year, which concluded on January 26, the revenue from China amounted to $17 billion, accounting for approximately 13% of the firm's overall sales.
However, other news outlets have highlighted that some of these chips keep ending up in China despite efforts to prevent this. Additionally, Nvidia has stated publicly that it lacks the ability to monitor where its products go once they are sold.
U.S. Representative Bill Foster, a Democrat representing Illinois who previously had a career as a particle physicist, stated that the means to monitor chips post-sale are easily accessible, noting that much of this capability is already integrated into Nvidia’s processors. Technical specialists corroborated these views when consulted.
During his distinguished scientific career, Foster developed numerous computer chips. He intends to propose legislation in the upcoming weeks that will instruct U.S. regulators to establish guidelines for two critical aspects: monitoring tracking chips to confirm their presence only in locations permitted by export control licenses, and ensuring these chips cannot operate unless correctly licensed according to export regulations.
Foster mentioned that there are already reliable accounts—some of which haven’t been made public—indicating widespread chip smuggling activities.
"This is not an imaginary future problem," Foster told . "It is a problem now, and at some point we're going to discover that the Chinese Communist Party, or their military, is busy designing weapons using large arrays of chips, or even just working on (artificial general intelligence), which is as immediate as nuclear technology."
Nvidia declined to comment for this story.
Chip smuggling has taken on new urgency after the emergence of China's DeepSeek, whose AI systems posed a strong challenge to U.S. systems and were built with Nvidia chips that were prohibited for sale to China, according to analyst firm SemiAnalysis. Prosecutors in Singapore have charged three people, including one Chinese national, with fraud in a case that involved servers that may have contained Nvidia chips.
Although this technology hasn't been widely implemented yet, the capability to confirm the whereabouts of chips already exists. Two insiders familiar with their processes say that Alphabet’s Google utilizes the tracking of locations for its proprietary AI chips and other components within its extensive array of data centers for safety reasons.
Google did not reply to requests for comments.
Foster’s bill would mandate the U.S. Department of Commerce to develop regulations within half a year necessitating the implementation of this technology.
BIPARTISAN SUPPORT
Foster's proposal has backing from other Democrats like Representative Raja Krishnamoorthi, who leads the House Select Committee on China. "Chip-based location verification presents an innovative approach we ought to consider for preventing these illegal activities," stated Krishnamoorthi.
Although Republicans back the idea, they haven’t endorsed particular bills since such proposals haven’t been presented yet. As the chairperson of the committee, Congressman John Moolenaar endorses the notion of location monitoring and intends to discuss possible legal strategies with legislators from both the House and the U.S. Senate later this week.
Moolenaar stated that the Select Committee strongly supports bipartisan efforts to mandate that firms such as Nvidia integrate location tracking capabilities into their advanced AI processors, noting that the necessary technology is already available.
The system for confirming chip locations would depend on these chips sending data to a secure computer server. The validation process would measure how long the signal takes to arrive at the server, based on the principle that electronic signals travel at the speed of light. This method hinges on understanding that communication signals propagate at this fixed velocity.
Tim Fist, who previously worked as an engineer and directed emerging technology policy at the Washington-based think tank Institute for Progress, stated that this kind of monitoring could offer a broad, nation-wide position indicator for the chips. However, this level of detail represents significantly more data than what the Bureau of Industry and Security—the division within the U.S. Commerce Department tasked with enforcing export regulations—currently possesses.
Fist mentioned that BIS hasn’t figured out which specific computer chips should be considered a top priority for examination after being exported abroad. However, with improved location verification, “they can now categorize the array of chips available globally into those that are almost certainly not illicitly obtained and those that require closer scrutiny,” he explained.
Foster's second legislative goal of preventing AI chips from booting up if they are not properly licensed under U.S. export controls would be more technologically complex to implement than location verification, but he said the time has come to begin discussions for both efforts.
"We've gotten enough input that I think now we can have more detailed discussions with the actual chip and module providers to say, 'How would you actually implement this?'" Foster told .
(Reported by Stephen Nellis and Max A. Cherney from San Francisco; Edited by Kenneth Li, Matthew Lewis, and Kate Mayberry)