Photo | Future Publishing | Getty Images
Nvidia The company is developing software that can provide location verification for its AI graphics processing units (GPUs), and the move comes as the U.S. government ramps up efforts to block the use of restricted chips in countries such as China.
The company said in a blog post Wednesday that the opt-in service uses a client software agent that Nvidia chip customers can install to monitor the health of their AI GPUs.
Nvidia also said that customers will be able to “visualize their GPU fleet usage globally or by compute zone (a group of nodes registered to the same physical or cloud location) in a dashboard.”
However, NVIDIA told CNBC in a statement that the latest software does not give it or any outside party the ability to disable the chip.
“There’s no kill switch,” he added. “There is no ability for NVIDIA to remotely control or take any action on registered systems regarding GPU health. This is read-only telemetry sent to NVIDIA.”
Telemetry is an automated process that collects data from remote or inaccessible sources and sends it to a central location for monitoring, analysis, and optimization.
The ability to locate a device depends on the type of sensor data collected and transmitted, such as IP-based network information, timestamps, and other system-level signals that can be mapped to physical or cloud locations.
A screenshot of the software posted on Nvidia’s blog showed details such as the machine’s IP address and location.
A screenshot of the software posted on Nvidia’s blog showed details such as the machine’s IP address and location.
Screenshot from Nvidia Blog | Opt-in NVIDIA Software Enables Data Center Fleet Management
Lukasz Olejnik, a senior research fellow at King’s College London’s School of War Studies, said NVIDIA suggested its GPUs do not have hardware tracking technology, but the blog did not specify whether the data was “using customer input, network data, cloud provider metadata, or other methods.”
“In principle, the transmitted data also contains metadata, such as network addresses, which could actually enable location determination,” Olejnik, who is also an independent consultant, told CNBC.
The software could also detect unexpected usage patterns that differ from those declared, he added.
Nvidia’s latest feature follows a request from the Washington state legislature to include tracking software on its chips that could help enforce export controls.
These rules prohibit Nvidia from selling its more advanced AI chips to companies in China and other prohibited regions without a special license. President Trump recently said he plans to lift some of these export restrictions, but restrictions on Nvidia’s cutting-edge chips will remain in place.
In May, Sen. Tom Cotton and a bipartisan group of eight senators introduced the Chip Security Act. If passed, the bill would require advanced AI chips to have security mechanisms and location verification.
“Companies affected by U.S. export controls or China-related regulations could use this system to verify and prove that their GPU fleets remain in approved locations and states, and potentially demonstrate compliant usage to regulators,” Olejn noted.
“It can actually help with compliance and indirectly make investment prospects more positive.”
Pressure on NVIDIA has increased following a Justice Department investigation into a suspected smuggling ring that transported more than $160 million worth of NVIDIA chips to China.
But Chinese authorities pushed back, warning Nvidia not to include tracking capabilities or “potential backdoors or vulnerabilities” in its chips.
Following a national security investigation into some of Nvidia’s chips to check for these backdoors, Chinese authorities have blocked local tech companies from purchasing products from American chip designers.
Despite US President Donald Trump giving Nvidia the green light to ship previously restricted H200 chips to China, the Chinese government is reportedly undecided on whether to allow imports.
