DOE's LBNL selects Dell-Nvidia system for NERSC-10 'Doudna' supercomputer
Built with AI in mind
Built with AI in mind
Test was carried out by data center operator Aquatron in partnership with the Los Alamos National Laboratory
But says it would have made billions more if not for Chinese restrictions
Systems have been plagued by overheating, software bugs, and leaking liquid cooling, engineers claim
Comes after US government imposed further restrictions on chipmaker’s H20 chips
Chipmaker will also establish AI technology center in the country
Will be housed in the Abilene, Texas, data center campus
New architecture designed to support 1MW+ racks in AI data centers
Would be chipmaker’s first investment in quantum company developing physical hardware
Companies claim to be targeting 'largest AI campus in Europe'
First phase of deployment will double UPenn’s processing capacity
Cloud giant launches P6-B200 Elastic Compute Cloud instances
Nvidia DGX Cloud Lepton gives users access to GPUs from a global network of cloud providers
Looks to be home to an OpenAI Stargate deployment
Fujitsu and Qualcomm will be first companies to integrate NVLink Fusion into their chips
System will contain 10,000 Nvidia Blackwell Ultra GPUs
B200s now generally available
$10bn agreement is separate from Nvidia GPU deal
Along with similar AMD deal
Oracle deploys "thousands" of Blackwell GPUs
AI cloud firm is less than a year old
Will be hosted in Italy
Chip launch follows additional US export restrictions on Nvidia H20 GPUs to China
Follows Nvidia's announcement it plans to produce AI supercomputers in the US