One other week, one other slew of Nvidia AI offers with Asian companions.
Final week Nvidia boss Jensen Huang unveiled a sequence of partnerships on a go to to Taipei, together with with SKT on bodily AI.
Now he is in Seoul asserting a number of offers with Korean corporations, notably with the SK Group, the mum or dad of SK Telecom and long-time Nvidia provider SK Hynix.
The dimensions and breadth of those tie-ups underscore the significance of Korea not solely to Nvidia’s international provide chain but additionally as a spot that may validate its tech at scale.Â
In a joint briefing with Huang at SK headquarters on Monday, SK Group chairman Chey Tae-won stated the brand new agreements signaled a deeper alliance that goes past simply reminiscence chips, Korea Herald reported.
“Most of our cooperation thus far has been centered on reminiscence, however any more, we are going to increase the partnership to the SK Group stage,” Chey stated.
In probably the most vital announcement, SK Hynix and Nvidia struck a multi-year deal for reminiscence provide and joint R&D on next-generation reminiscence and AI structure.
AI cloud
Nvidia stated the partnership would allow reminiscence provide “to maintain tempo with Nvidia’s infrastructure roadmap and the sustained buildout of AI infrastructure worldwide.”
“Collectively, we’re codeveloping the subsequent technology of reminiscence for AI factories and making use of AI to how we design and manufacture semiconductors,” Chey stated.
SK Telecom’s position within the enlarged partnership is to construct a gigawatt AI cloud, anticipated to come back on-line in 2027, Nvidia stated.Â
The GPU-based cloud is aimed toward serving rising demand for AI infrastructure and will probably be constructed on Nvidia’s DSX platform, designed to provide low-cost tokens at excessive power effectivity.
It should help coaching, inference and agentic workloads, together with sovereign AI, initially for South Korea however with plans to increase to elsewhere in Asia.Â
Nvidia additionally introduced AI infrastructure offers with Naver and the LG Group.
Naver, whose companies vary from search to purchasing and cloud, will use Nvidia chips to develop its sovereign AI infrastructure, beginning at 55MW however with the intention of increasing to gigawatt scale.
It should additionally use Nvidia chips to speed up its information heart enterprise and to spice up its sovereign AI capability in Europe and the Center East.
Nvidia’s partnership with LG covers a number of rising applied sciences together with robotics, autonomous driving, information facilities and cloud.
The AI manufacturing facility will present LG Group with accelerated computing infrastructure to coach and validate AI-based functions throughout the corporate, Nvidia stated.
