The South Korean tech powerhouse, Samsung, announced on Thursday that it intends to acquire and utilize a batch of 50,000 Nvidia graphic processing units to enhance its production of chips for mobile gadgets and robots.
This batch of 50,000 Nvidia GPUs is earmarked for a facility that Samsung is designating as an “AI Megafactory.” However, Samsung has not disclosed when this facility will come into existence.
This marks the latest noteworthy collaboration for Nvidia, whose chips continue to be pivotal in the construction and operation of cutting-edge artificial intelligence technologies.
The partnership with Samsung follows Nvidia CEO Jensen Huang’s declaration on Tuesday in Washington, D.C., regarding Nvidia working with firms such as Palantir, Eli Lilly, CrowdStrike and Uber.
Following this address, Huang was seen in South Korea enjoying beer with Samsung Chairman Lee Jae-yong and other industry leaders, as reported by local media. Other South Korean firms, including SK Group and Hyundai, are also putting similar quantities of GPUs to use, as per Nvidia.
“We are collaborating closely with the South Korean government to back its ambitious plans in AI,” stated Raymond Teh, Nvidia’s senior vice president for Asia-Pacific, during a call with reporters on Wednesday.
These partnerships reinforce Huang’s assertion made Tuesday that Nvidia’s business portfolio amounts to $500 billion from its current generation GPU, named Blackwell, in addition to its forthcoming GPU, named Rubin.
This projection has contributed to a rise in Nvidia’s stock, establishing the company as the first to achieve a market valuation of $5 trillion.
On Thursday, Nvidia representatives communicated that they will collaborate with Samsung to adjust the Korean firm’s chipmaking lithography platform in order to accommodate Nvidia’s GPUs. This transformation is anticipated to enhance Samsung’s performance by 20 times, according to Nvidia’s representatives. Samsung will additionally utilize Nvidia’s simulation software known as Omniverse. Recognized for its mobile devices, Samsung also indicated it would employ the Nvidia GPUs to execute its own AI models for its products.
Beyond serving as a partner and customer, Samsung is also a critical supplier for Nvidia.
Samsung produces the type of high-performance memory that Nvidia utilizes extensively, alongside its AI processors, referred to as high bandwidth memory. Samsung announced plans to collaborate with Nvidia to modify its fourth-generation HBM memory for application in AI chips.
WATCH: An evening in Seoul: Nvidia, Samsung, and Hyundai executives connect over fried chicken and soju
