Samsung has unveiled plans to purchase and deploy a cluster of 50,000 Nvidia graphic processing units (GPUs), a major move intended to enhance its semiconductor manufacturing for mobile devices and robotics.
The facility, which Samsung has dubbed an “AI Megafactory,” will leverage these GPUs to bolster its artificial intelligence capabilities in chip production. While Samsung has yet to provide specifics on the construction timeline or operational details of the facility, the announcement signals the company’s continued push into AI-powered manufacturing.
This move adds to Nvidia’s growing list of high-profile collaborations, underscoring the ongoing demand for its industry-leading chips in driving advanced AI. The partnership with Samsung follows recent remarks by Nvidia CEO Jensen Huang, who, during an address in Washington, D.C. on Tuesday, announced collaborations with firms such as Palantir, Eli Lilly, CrowdStrike, and Uber.
According to local media reports, Huang was later seen in South Korea with Samsung Chairman Lee Jae-yong and several business executives—a testament to Nvidia’s deepening ties in the region. Nvidia also noted that major Korean firms, including SK Group and Hyundai, are undertaking similar GPU deployments.
Nvidia’s senior vice president for Asia-Pacific, Raymond Teh, emphasized the company’s commitment to Korea’s AI ambitions, stating on a media call Wednesday that Nvidia will work closely with the Korean government to support its AI leadership ambition.
These strategic partnerships align with Huang’s assertion that Nvidia’s current and next-generation GPUs—Blackwell and Rubin—are propelling a $500 billion business pipeline. The company’s bullish outlook has played a key role in Nvidia’s stock performance, recently lifting its market capitalization to $5 trillion, the first ever for a chipmaker.
Beyond the procurement of GPUs, Nvidia representatives stated on Thursday that the company is partnering with Samsung to adapt its chip fabrication process, specifically Samsung’s lithography platform, for Nvidia GPU integration. This customized process is projected to deliver 20-fold performance improvements for Samsung’s manufacturing.
Additionally, Samsung will deploy Nvidia’s Omniverse simulation software and run proprietary AI models on the new hardware to advance its mobile technology.
Samsung is not only a customer but also a critical supplier for Nvidia. The Korean conglomerate provides the high-performance memory—known as high bandwidth memory (HBM)—that Nvidia uses in large-scale AI chip deployments. Samsung indicated it will collaborate with Nvidia to refine its fourth-generation HBM memory for future AI chip applications.
