Nvidia now reportedly requires cloud providers in Southeast Asia to enroll in its Fleet Intelligence tracking software before they can become official Nvidia Cloud Partners, according to people familiar with the new requirements. Nvidia did not immediately comment.
The partner badge is worth chasing. It brings Nvidia's public endorsement and preferred access to its scarce AI chips, which hands the company leverage without any law compelling anyone.
Fleet Intelligence collects data on Nvidia hardware across a customer's infrastructure, including the location of individual accelerators. It launched in December 2025 as a data center management tool. Outside the partner requirement, participation is still voluntary, and Nvidia does not automatically receive location data from every user. The company has not said whether existing partners must enroll to keep their status, or whether the rule will extend beyond the region.
The target is diversion. The US bars Chinese companies from buying advanced AI accelerators without a license, yet authorities believe billions of dollars of them have reached Chinese buyers through third parties. The nonprofit C4ADS examined 2022 to 2025 transactions and found $13.4 million of A100, H100 and AD102-series GPUs moving through regional routes. It also linked $4.6 billion in hardware purchases to Singapore-based Megaspeed International, which has been under US and Singaporean investigation for more than a year. Nvidia CEO Jensen Huang called his dealings with the firm "completely normal"; Megaspeed denied wrongdoing.
The Super Micro case shows why paperwork is a weak control. US prosecutors accused people associated with the server maker of helping divert about $2.5 billion of Nvidia-powered servers to China, using falsified documents and dummy server shells, with real hardware shipped onward after inspections. Software that reports where a chip is running is harder to stage than a document trail, though Nvidia's earlier telemetry and network latency approach only estimates location.
This fits a pattern. Nvidia has also reportedly removed half of the Asian companies from its pre-approved buyer list and told cloud providers to keep at least 51% of their computing services with customers outside China.
The weak point is the incentive. A badge only reaches companies that want one, and a buyer planning to reroute chips has little reason to apply.