Nvidia will integrate an internal GPS into its AI chips to monitor their route and final destination

Nvidia is developing technology that will allow it to determine the country where its artificial intelligence chips are operating, a measure aimed at preventing these processors from reaching territories where their export is restricted. The company has privately demonstrated this tool to partners and customers, although it has not yet officially launched it. The feature would operate as optional software that users could install in their data centers.

According to Reuters, the technology utilizes confidential computing capabilities built into the latest generation GPUs and would allow the location of each chip to be estimated by analyzing the communication time with Nvidia's servers. This mechanism would give companies an additional way to monitor the whereabouts of their processors, a crucial practice at a time when control over AI semiconductors has become a geopolitical issue.

This initiative comes in a context marked by increasing scrutiny from the U.S. government, which seeks to prevent advanced chips from reaching China or other countries subject to restrictions. In recent months, the Department of Justice has opened investigations into attempts to smuggle Nvidia processors valued at more than $160 million, allegedly destined for the Chinese market.

A System Designed for Monitoring and Operational Security

In a statement, Nvidia confirmed that it is working on a new software service for data center operators. This "agent," installed by the customer, will collect GPU telemetry to monitor the health, integrity, and inventory of fleets of chips dedicated to artificial intelligence. The company explains that the location verification is not intended as a surveillance mechanism, but rather as an additional function within this monitoring system.

The feature will debut in the Blackwell chips, the company's latest line. These processors incorporate security enhancements compared to the Hopper and Ampere generations, making it easier to implement attestation processes—mechanisms that guarantee the hardware is running as designed. Nvidia is also evaluating extending this functionality to previous generations.

Among the customers who could benefit from the software are operators of large data centers and companies that manage critical infrastructure. These organizations typically manage thousands of GPUs in different locations around the world, so a tool that confirms their location and operational status represents a valuable element of internal control.

Regulatory Tensions Between the United States and China

The ability to verify the location of the chips could address demands from U.S. lawmakers, who have insisted on stricter measures to prevent China from accessing advanced processors. Both the White House and members of Congress from different parties have called for mechanisms to prevent the diversion of these devices.

However, the development of this feature has also raised concerns in China. The country's main cybersecurity regulator reportedly summoned Nvidia to request explanations, amid fears that this technology could serve as a remote access point for the U.S. government. The company has rejected this idea, asserting that there are no backdoors in its products.

These concerns intensified after U.S. President Donald Trump recently stated that he would allow sales of the H200 chip, a direct predecessor of the Blackwell family. Even so, international policy experts consider it uncertain whether China will authorize the import of these models under the current regulatory conditions.

A launch that could set a precedent

Although Nvidia has not yet set a date for the launch of its verification technology, its mere existence suggests a potential standard in the semiconductor industry. Location tracking could become a common tool in a global context where AI chips are considered strategic assets.

Software specialists indicate that it is technically feasible to implement this function without compromising user security or exposing their data. The company, for its part, maintains that the system's primary objective is to improve operational management and provide traceability for customers who handle enormous volumes of distributed hardware.