Samsung to buy 50,000 Nvidia GPUs to automate chip manufacturing
Jensen Huang, chief executive officer of Nvidia Corp., during the keynote speech at the Nvidia GTC (GPU Technology Conference) on Tuesday, October 28, 2025 in Washington, DC, USA.
Kent Nishimura | Bloomberg | Getty Images
Korean semiconductor giant Samsung said on Thursday it plans to buy and distribute a 50,000-strong cluster. Nvidia graphics processing units to improve chip manufacturing for mobile devices and robots.
The 50,000 Nvidia GPUs will be used to create a facility that Samsung calls the “AI Mega Factory.” Samsung did not provide details on when the facility would be built.
It’s the latest flashy partnership from Nvidia, whose chips remain essential to building and deploying advanced artificial intelligence.
The collaboration with Samsung comes after Nvidia CEO Jensen Huang announced in Washington DC on Tuesday that Nvidia is selling collaborations with the following companies: palantir, Eli Lilly, CrowdStrike And Uber.
Shortly after the speech, Huang was seen drinking beer with Samsung Chairman Lee Jae-yong and other business leaders in South Korea. According to local media. Nvidia said other Korean companies, including SK Group and Hyundai, are using similar amounts of GPUs.
“We are working closely with the Korean government to support their ambitious leadership plans in artificial intelligence,” Raymond Teh, Nvidia’s senior vice president for Asia-Pacific, told reporters on Wednesday.
The partnerships support Huang’s claim on Tuesday that Nvidia has a combined $500 billion book of business from its current-generation GPU called Blackwell, in addition to its next-generation GPU called Rubin.
The forecast helped Nvidia’s shares soar, and the company became the first company to reach a $5 trillion market cap.
Nvidia representatives on Thursday said the Korean company will work with Samsung to adapt its chipmaking lithography platform to work with Nvidia’s GPUs. Nvidia representatives said this process will result in 20 times better performance for Samsung. Samsung will also use Nvidia’s simulation software called Omniverse. Samsung, known for its mobile phones, also said that it will use Nvidia chips to run its own artificial intelligence models on its devices.
In addition to being a business partner and customer, Samsung is also a major supplier to Nvidia.
Along with AI chips, Samsung also makes so-called high-bandwidth memory, a type of high-performance memory that Nvidia uses in large quantities. Samsung said it will work with Nvidia to tune fourth-generation HBM memory for use in AI chips.
WRISTWATCH: Night in Seoul: Nvidia, Samsung and Hyundai bosses bond over fried chicken and soju
