Samsung’s HBM3E chip, top, and other DDR modules
Samsung’s HBM3E chip recently won the approval of Nvidia chief Jensen Huang © SeongJoon Cho/Bloomberg

Samsung Electronics will produce the latest generation of semiconductors in the US two years ahead of its rival Taiwan Semiconductor Manufacturing Company, in a further boost to President Joe Biden’s efforts to bring advanced chip production on to home soil.

US commerce secretary Gina Raimondo announced on Monday that the South Korean chipmaker was set to produce chips at the 2 nanometre level of miniaturisation at a new fabrication plant, or fab, it was building in the city of Taylor in Texas. It will be part of a $40bn investment made by the company in capabilities ranging from the manufacturing of microprocessors to advanced chip packaging and research and development work.

Samsung’s investment will be topped up with as much as $6.4bn in direct funding received under the US Chips and Science Act, a week after the US government announced that TSMC would receive up to $6.6bn from the flagship subsidy programme to support its chip expansion plans in Arizona.

The first of Samsung’s new Taylor fabs will start making 2nm chips in 2026, according to a senior US official. TSMC is set to produce 2nm chips at an Arizona facility from 2028.

The Biden administration aims to increase domestic production of advanced chips from zero to 20 per cent of the world’s supply by the end of the decade, amid concerns that they could be cut off by a natural disaster or possible future conflict in east Asia.

“Much of the semiconductor supply chain . . . is concentrated in a couple of Asian locations, and that leaves the US supply chain incredibly vulnerable to disruption,” said Raimondo, adding that the new Samsung investment “puts us on track to hit our [20 per cent] goal”.

“We are now making these investments which will allow the United States to once again lead the world, not just in semiconductor design . . . but also in manufacturing, advanced packaging and research and development.”

An initial $17bn investment, part of a planned $40bn in total, was announced by Samsung in 2021 to build its first fab in Taylor, which will now make 2nm and 4nm chips.

The new capital spending adds a second fab that will also make 2nm and 4nm chips, as well as the construction of an advanced chip packaging facility for “2.5D packaging” of processor and memory chips.

Advanced packaging is a crucial stage in the production of artificial intelligence chips such as Nvidia’s H100, which is used to train generative AI systems such as OpenAI’s ChatGPT.

TSMC, the world’s biggest contract chipmaker, carries out 2.5D packaging for Nvidia’s most powerful chips. But it currently has no plans for such an advanced packaging facility in the US.

“Right now, even chips that are made in the US are still shipped in many cases to Taiwan to be packaged, including the chips used in defence systems,” said Raimondo.

With its new plans for Taylor, Samsung will be able to combine the latest graphics processing units and its own high-bandwidth memory chips in advanced AI products packaged at its US facility.

Last month, Nvidia chief executive Jensen Huang wrote “JENSEN APPROVED” on a Samsung HBM3E chip on display at a developer conference in California, prompting speculation that Samsung was on course to secure the leading AI chip vendor as a customer.

“To meet the expected surge in demand from US customers for future products like AI chips, our [Taylor] fabs will be equipped for cutting-edge process technologies and help advance the security of the US semiconductor supply chain,” said Kye Hyun Kyung, president and chief executive of Samsung’s chip division, on Monday.

Lael Brainard, the White House economic adviser, said Samsung had also made a “series of commitments to enable it to manufacture chips directly for the Department of Defense”.

The senior US official also noted that Samsung’s R&D fab for developing future generations of chips would be only the fourth such facility in the world and the first built in the US by a non-US company.

Monday’s announcement was the sixth and final disbursement from the first round of Chips Act grants. US chipmaker Intel has received the largest sum, up to $8.5bn, while GlobalFoundries, Microchip Technology and BAE Systems have also received funding.

Additional reporting by Lauren Fedor in Washington and Michael Acton in San Francisco

Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content (opens in new window) CommentsJump to comments section

Follow the topics in this article

Comments