Southeast Asia computing providers must now register with Nvidia's Fleet Intelligence tracking system to achieve official Nvidia Cloud Partner status, according to a Bloomberg report. The requirement aims to help the chipmaker monitor the deployment of AI accelerators as U.S. authorities investigate the smuggling of restricted hardware into China through regional networks. While Fleet Management was introduced last December as a data center management tool to monitor hardware health and inventory, the new mandate allows Nvidia to gather information regarding the specific locations of individual accelerators within a provider's infrastructure.
Participation in the tracking system has previously been voluntary, and Nvidia does not automatically receive location data from all customers using the software. The company has not yet clarified if existing Cloud Partners must enroll to maintain their designation or if the mandate will expand beyond Southeast Asia. This tracking capability follows reports that Nvidia has developed software-based GPU tracking using telemetry and network latency measurements to estimate the operating locations of processors.
U.S. authorities have expressed frustration over billions of dollars worth of hardware reaching Chinese buyers through third-party networks, with Southeast Asia serving as a primary transit point. A C4ADS investigation identified $13.4 million worth of Nvidia A100, H100, and AD102-series GPUs moving through regional trade routes between 2022 and 2025. The nonprofit also linked $4.6 billion in hardware purchases to Singapore-based Megaspeed International, a firm currently under investigation by U.S. and Singaporean authorities for suspected chip diversion.
Nvidia CEO Jensen Huang defended the company's dealings with Megaspeed, which has denied any wrongdoing, describing the transactions as “completely normal,” adding, “we’ve studied all their deployments of our GPUs and counted every one.” The scrutiny follows allegations from U.S. prosecutors that individuals associated with Super Micro helped divert approximately $2.5 billion worth of Nvidia-powered servers to China using falsified paperwork and dummy server shells.
