Samsung Texas plant to make HBM chip parts for AMD, Nvidia

Samsung Texas plant to make HBM chip parts for AMD, Nvidia
Samsung powers AI chips

Samsung Electronics is expanding the role of its Taylor, Texas, semiconductor investment to support artificial intelligence memory supply chains. The company’s second foundry plant there is set to produce base dies for high-bandwidth memory chips for AMD and Nvidia, with most secured volume reportedly tied to AMD.

Highlights

  • Samsung will use its new Taylor, Texas plant to produce HBM chip base dies for AMD and Nvidia, with construction starting this month and output mostly committed to AMD.
  • The facility will span 2.7 million square feet, 1.5 times larger than Samsung’s first Taylor plant, and employ advanced 2-nanometre technology.
  • Equipment installation is set for February 2029 with mass production in 2030, reinforcing Samsung’s strategic semiconductor role in the U.S. amid strong AI-related memory demand.

Taylor plant plans and production timeline

As reported by Reuters, citing Maeil Business Newspaper, Samsung plans to use its second foundry plant in Taylor, Texas, to manufacture base dies for high-bandwidth memory, or HBM, chips for AMD and Nvidia.

Construction is due to start formally this month, and the plant has reportedly secured orders for its full output. Most of that committed volume is for AMD, according to the report.

The facility is set to use advanced 2-nanometre technology and span about 2.7 million square feet. Maeil said that makes it roughly 1.5 times the size of Samsung’s first Taylor plant, which has a contract with Tesla.

Equipment installation is expected to begin in February 2029, with mass production targeted for 2030. Samsung Electronics declined to comment on the report.

Implications for U.S. chip manufacturing

The reported production plan points to a deeper Samsung role in the U.S. semiconductor manufacturing base, particularly in components tied to AI-related memory demand. Base dies are a key part of HBM chips, which are widely used in advanced computing workloads.

For Taylor, the project adds to the strategic importance of the Texas site as chipmakers and their customers seek more advanced manufacturing capacity in the U.S. The reported order backlog also suggests early customer commitment ahead of construction and equipment installation.

Our earlier article on the U.S. AI data center buildout highlighted how soaring funding costs, construction delays and grid constraints are starting to slow projects and test investor confidence. We noted that only a minority of planned developments are viewed as truly feasible, as developers face higher capital needs, community pushback and pressure to secure reliable power.

This material may contain third-party opinions, none of the data and information on this webpage constitutes investment advice according to our Disclaimer. While we adhere to strict Editorial Integrity, this post may contain references to products from our partners.