Nvidia links Southeast Asia cloud partner status to chip tracking
The requirement is intended to widen participation in a voluntary monitoring programme amid concerns over chip smuggling to China, Bloomberg reported.
Customers seeking Nvidia Corp’s official cloud partner designation in Southeast Asia must now join a system that monitors its artificial intelligence (AI) hardware, Bloomberg reported, citing people familiar with the matter.
The requirement concerns Nvidia’s fleet intelligence system, which records deployment information and the geographical location of individual components. Recognition as Nvidia Cloud Partners gives companies preferential chip access as well as public endorsement from the chipmaker.
People familiar with the measure said Nvidia aims to attract more participants to the otherwise voluntary programme by linking it to partner recognition. The step forms part of efforts to curb the diversion of advanced chips to Chinese companies, whose purchases generally require a licence under US restrictions.
Participation does not, by default, require location records to be shared with either Nvidia or US authorities. A company spokesperson previously told Bloomberg that regulators in Washington would contact users directly if they wanted information held in the database.
Nvidia did not say whether current partners would need to join the programme to keep their designation. It also declined to address whether the condition would be introduced elsewhere.
In another measure, Nvidia has started telling customers that rentals to Chinese companies should not exceed half of their Nvidia computing power. US policy generally allows Chinese businesses to access accelerators operated by others outside China.
