THE ESSENTIALS IN 30 SECONDS

  • Bloomberg reports that Southeast Asian cloud providers seeking Nvidia Cloud Partner status must join Fleet Intelligence, which logs the physical location of each accelerator.
  • The service, introduced last December, is otherwise opt-in; Nvidia has said it provides read-only telemetry and includes no kill switch.
  • Nvidia did not comment on the report, and it is unclear whether existing partners or other regions are covered.

Nvidia has started requiring cloud operators in Southeast Asia to enrol in its Fleet Intelligence service if they want to be recognised as Nvidia Cloud Partners, Bloomberg reported on November 6, citing people familiar with the matter. The report was relayed by Tom's Hardware and AI Weekly; Nvidia was not available for comment, and the company has not confirmed the policy publicly.

What the requirement involves

The Nvidia Cloud Partner label signals the chipmaker's public endorsement and, according to the reports, gives preferred access to its most sought-after accelerators. Fleet Intelligence, introduced last December as a data-centre management tool, gathers operational data on Nvidia hardware across a participating company's infrastructure, including the physical location of individual chips. Outside this new condition, joining remains voluntary, and Nvidia does not receive location data from customers that have not signed up.

When the service was first described in December 2025, The Register reported that its main purpose was streaming health telemetry to reduce failures, with location checks as a secondary feature. As the outlet understood it, the location check resembles a ping whose response time is compared with expected ranges, relying on security features already present in GPUs since the Hopper generation. An Nvidia spokesperson told the outlet that "there is no kill switch" and that the telemetry is read-only.

Tom's Hardware notes that the reports do not say whether existing partners must enrol to keep their status or whether the rule could extend beyond Southeast Asia.

Why Southeast Asia

US rules bar Chinese companies from buying advanced AI accelerators without a licence, yet US officials believe billions of dollars of restricted hardware has reached China through intermediaries, with Southeast Asia a key transit point. Earlier this year, US prosecutors accused people linked to Super Micro of diverting about $2.5 billion in Nvidia-based servers to China through regional middlemen.

The step adds to other pressure Nvidia has reportedly applied to its customers in 2026, including trimming its list of pre-approved Asian buyers and asking cloud providers to keep at least 51% of their business outside China. Location tracking goes further by giving Nvidia a technical record of where chips actually run, rather than relying only on contracts and declarations. How reliably latency-based estimates can detect diverted hardware has not been independently assessed.