←News detail

Rohan PaulUpdated Nov 5 23:09

Moonshot AI seeks to rent Nvidia Blackwell chips in European data centers

AI summary

CNBC reported that Moonshot AI, the Chinese lab behind Kimi, has held talks with European non-Chinese GPU rental providers known as neoclouds to rent Nvidia Blackwell chips, with one cluster exceeding 8,000 units, primarily to provide Kimi services to users. A proposed lease shows that Moonshot is willing to prepay 30% to 50% more than non-Chinese tenants to buffer neoclouds against the risk of Washington banning remote access. U.S. export controls prohibit shipping these chips to China but do not broadly cover renting servers from overseas; Nvidia is said to require cloud providers to allocate at least 51% of services in new contracts to non-Chinese customers, with the remaining 49% still available to Chinese buyers.

Why read this

AI score 71

The original report disclosed the specific negotiation terms and premium rates for Moonshot's rental of Blackwell in Europe, revealing an alternative path for obtaining computing power under export controls.

Topics

Tags

Body

CNBC: Moonshot AI, the Chinese lab behind Kimi, has sought to rent Nvidia Blackwell chips in European data centers.

It discussed deals with neoclouds, non-Chinese GPU rental firms that can buy the hardware.

U.S. export controls bar shipping those chips to China but don’t broadly cover renting them from servers abroad.

The White House accused Moonshot in July of using the same route in Thailand, and a source said the European talks came afterward.

One person described clusters of 8,000 or more Blackwell GPUs, sought mainly for serving Kimi to users.

Under one proposed lease, Moonshot would pay 30% to 50% more upfront than a non-Chinese renter, cushioning neoclouds if Washington bans remote access.

Nvidia reportedly told cloud providers that on new contracts, at least 51% of services must go to non-Chinese customers.

That cap still leaves 49% for Chinese buyers, letting a large European neocloud rent Moonshot 8,000 GPUs.

Source

Rohan Paul