Samsung Electronics will dedicate its second semiconductor plant in Taylor, Texas, to manufacturing base dies for high-bandwidth memory chips destined for AMD and Nvidia. Industry sources reported this week that the South Korean company has already booked orders covering the facility’s total planned capacity ahead of ground breaking.

AMD accounts for the majority of the secured volume at the site, where formal construction is scheduled to begin this month.

The plant will cover about 2.7 million square feet, roughly 1.5 times the size of Samsung’s first Taylor facility, which holds a supply contract with Tesla, and will deploy an advanced 2-nanometre manufacturing node.

Equipment installation will start in February 2029. Commercial mass production will follow in 2030.

Samsung Electronics declined to comment on the customer allocations and production timelines. The company previously dedicated its initial Taylor plant to automotive and computing contracts, including work for Tesla.

Base dies form the foundation of high-bandwidth memory stacks, acting as the logic layer that links dynamic random-access memory directly to high-performance graphic processing units. Securing dedicated foundry capacity for these components removes a critical choke point for server makers, data platform operators and consumer tech hardware vendors that rely on fast graphic processors.

Committing the plant entirely to external graphic chip designers locks Samsung into heavy capital deployment years before revenue flows. Cleanroom tools and extreme ultraviolet lithography systems must be ordered and amortised on schedules that leave little room for client order cancellations. The operational risk sits squarely on yield rates: running a 2-nanometre node at scale has challenged every major foundry in Asia, and transferring that production overseas amplifies tooling and labour risks.

Source link