Nvidia Ties Cloud Partner Status to Fleet Intelligence Tracking

Nvidia has reportedly introduced a new condition for computing providers in Southeast Asia. Firms that want to become official Nvidia Cloud Partners (NCPs) will have to enroll in a tracking system called Fleet Intelligence. According to people familiar with the matter, the step is meant to make it easier to monitor where the company's AI accelerators are being used.

The move comes as Washington grows increasingly concerned that restricted hardware is reaching China through regional smuggling networks. Nvidia has not yet responded to a request for comment on the matter.

Why NCP status matters and what the system collects

The NCP designation gives cloud providers Nvidia's public endorsement and priority access to in-demand AI chips. Those perks hand Nvidia considerable leverage over the companies involved. Firms seeking the status will now have to join Fleet Intelligence.

Fleet Management, the foundation of the system, was introduced last December primarily as a data center management tool. It gathers information about the Nvidia hardware in participating companies' infrastructure and also records the location of individual accelerators. That data could give Nvidia a clearer picture of whether chips are still operating after they leave the supply chain.

This is not, however, a blanket tracking mandate. Outside the NCP requirement, participation remains voluntary, and Nvidia does not automatically collect location data from every customer who uses the software. The company has also not said whether existing cloud partners must register to keep their status, or whether the requirement could later extend beyond Southeast Asia.

Smuggling networks and the Megaspeed file

Nvidia's work on locating its GPUs is not new. Last December, the company was reported to have developed software-based tracking that estimates where processors are running by using GPU telemetry and network latency measurements. Nvidia described the software as a customer-installed agent that monitors hardware health, integrity and inventory.

The US bans Chinese companies from buying advanced AI accelerators without a license. Even so, officials believe billions of dollars' worth of hardware has reached Chinese buyers through third-party networks, and Southeast Asia has emerged as a key transit point. An earlier review by a nonprofit identified several routes, including research initiatives, shell companies and shipments rerouted through other Southeast Asian countries. Examining transactions from 2022 to 2025, the group found $13.4 million in A100, H100 and AD102-series GPUs moving along regional trade routes. The same report links $4.6 billion in hardware purchases to Singapore-based Megaspeed International.

Megaspeed has become a particularly awkward case for Nvidia, which has been at the center of a probe by the US and Singapore for more than a year over suspected chip diversion. Nvidia CEO Jensen Huang defended the company's dealings with the firm last week, calling them "completely normal" and saying "we've reviewed every deployment of GPUs, we've counted every one of them." Megaspeed denied any wrongdoing. In April, Megaspeed's data center landlord in Malaysia terminated the contract and moved 68.4 megawatts of capacity to another cloud provider.

Scrutiny is tightening

Megaspeed is not the only example. Earlier this year, US prosecutors alleged that people linked to Super Micro helped divert roughly $2.5 billion worth of Nvidia-powered servers to China through Southeast Asian intermediaries. The scheme allegedly relied on forged paperwork and empty server shells, with the real hardware continuing on its way once inspections were passed.

Nvidia has also reportedly removed half of the Asian companies on its pre-approved buyer list and told cloud providers to reserve at least 51 percent of their computing services for customers outside China.