Copyright researchsnipers
At Samsung, 50,000 Nvidia GPUs will soon take over chip production. With the so-called “AI Megafactory” project, Samsung wants to significantly accelerate and optimize its production of processors for mobile devices and robotics. Comprehensive cooperation agreed The move marks another large-scale cooperation for Nvidia, whose chips are considered an indispensable basis for modern AI infrastructures. The deal was announced just two days after Nvidia CEO Jensen Huang appeared in Washington, where he unveiled new partnerships with several companies, including surveillance giant Palantir. According to South Korean media, Huang met shortly afterwards in Seoul with Samsung CEO Lee Jae-yong and other business leaders at an unofficial meeting, the US broadcaster reports CNBC. As part of this, the new cooperation was agreed. The company still left it open when the plant would actually go into operation. The agreements stipulate that Samsung and Nvidia will work together to adapt Samsung’s lithography platform to better suit the use of Nvidia chips. The goal is to ultimately increase production performance twenty-fold. Samsung will also use Nvidia’s Omniverse simulation software to develop its own AI models for smartphones and other devices. Nvidia on the road to success Nvidia should also benefit from the deal beyond GPU sales, because Samsung is not only a customer, but also an important supplier: The company produces the high-performance RAM (HBM) that is built into Nvidia’s AI chips. In the future, both companies want to work together on the further development of the next generation of memory HBM4 for use in AI applications. The partnerships fit into Nvidia’s expansive growth strategy: Huang had previously announced that the company currently has orders worth around $500 billion for its current Blackwell GPU generation and the upcoming Rubin series. This prospect helped Nvidia become the first company in the world to reach a market capitalization of $5 trillion.
