Samsung Reportedly Bags 3 nm Chip Orders from NVIDIA, Qualcomm, IBM, and Baidu
The orders would expand Samsung’s list of semiconductor clients, currently dominated by telecommunications companies, to high-performance computing majors.
Samsung already has four high-profile customers for its chips developed with the advanced 3nm process. Reportedly, NVIDIA, Qualcomm, IBM, and Baidu have chosen the South Korean hardware behemoth as their supplier for what are currently the most advanced semiconductors in the world.
According to people familiar with the matter who spoke with The Korea Economic Daily, Samsung, which counts Apple, Verizon, Deutsche Telekom, and Supreme Electronics as some of its biggest customers, has reportedly bagged American chip designers Qualcomm and NVIDIA, IT major IBM and Chinese AI/internet services company Baidu as its new customers.
Sources told KED that the four companies want to source chips from Samsung, which unlike Taiwan Semiconductor Manufacturing Company (TSMC), has already started mass production of the first generation of GAAFET- and 3 nm process-based semiconductors in Samsung’s Hwaseong foundry in June 2022.
So it would seem Samsung has got a head start in the race to become the supplier of advanced chips, and its goal to dethrone TSMC as the global semiconductor king (global market share more than thrice of Samsung’s ~17%) while keeping Intel, which fast-tracked chip development and fabrication last year, at bay.
“Some companies are cutting down on their transactions with Taiwanese companies. Instead, they are looking for a second and a third supplier in other countries such as Samsung,” an official at a back-end process firm told KED.
See More: TSMC To Kick Off 3nm Chip Production In the U.S. From Its $12B Arizona Chip Fab
Profitability from semiconductors is a numbers game, given its capital-intensive manufacturing process. So contracts from four big tech companies should certainly help Samsung narrow the gap with TSMC.
Samsung’s first generation 3 nm process delivers chips with up to 23% performance improvement, 45% higher power efficiency, and a wafer area reduction of 16% compared to its 5 nm node. The four companies are looking to leverage Samsung’s 3 nm node for respective markets, viz., GPUs (NVIDIA), smartphone and mobile devices (Qualcomm), CPUs (IBM), and AI-driven cloud data centers (Baidu).
The development also means the expansion of Samsung’s list of semiconductor clients, currently dominated by telecommunications companies, to high-performance computing majors.
The second generation 3 nm process-based chips, slated to go into production in 2023, can reduce power consumption by up to 50%, improve performance by 30% and reduce the wafer area by 35%. Samsung is also building a $17 billion chip fabrication plant in Williamson County, 40 miles from its existing facility in Austin, Texas.
The U.S. Commerce Department’s Bureau of Industry and Security (BIS) banned the export of 3 nm process technology, including GAAFET, in August 2022. Hence, it is highly probable but still unclear (given South Korea is a U.S. ally) that the second generation of 3 nm-based chips will be fabricated in the new Texas plant.
Qualcomm currently sources chips (Snapdragon 8 Gen 2 and Snapdragon 8+ Gen 1) produced on the 4 nm node from TSMC, while its Snapdragon 8 Gen 1 is based on Samsung’s 4 nm process.
Let us know if you enjoyed reading this news on LinkedIn, Twitter, or Facebook. We would love to hear from you!
Image source: Shutterstock