Sale occurred in fourth quarter of 2025

Company will also establish APAC’s largest Nvidia DGX Cloud cluster

Agreement represents the first “large-scale” Nvidia Grace-only deployment

Company says ChipStack AI Super Agent has already been deployed by Nvidia and Qualcomm

System will be operated by telco’s Viettel AI unit

Comes in same week Nvidia CEO downplays $100bn investment pledge in OpenAI

Government is reportedly weighing up conditions under which additional purchases will be greenlit

CPU had previously only been available as part of Vera Rubin Superchip

Aims to help speed up AI buildout

Bipartisan legislation passed the House 42-2

Comments made by Peng Xiao during an interview at Davos

Nvidia H200 and AMD MI325X GPUs will be subject to the levy

Exports will only be permitted if sufficient supply of the GPUs exists in the US

Meanwhile, Nvidia says upfront payments from Chinese customers purchasing H200s is not company policy

Next-gen Rubin platform will see DPUs help control all network I/O to and from the compute node

Combines Nvidia AI infrastructure with Lenovo’s hardware and hybrid AI factory services

Customers in China will also not be able to cancel, reconfigure placements, or ask for refunds

Expected to be available in the second half of 2026

Founder Jonathan Ross heads to Nvidia in deal that could be worth $20bn

Chips galore

Was tied to 10GW data center commitment

A victory for chipmakers and the Trump administration

Partners say agreement will support new market opportunities

Company calls report "factually incorrect"