Samsung's Second Texas Fab Report Points to AMD and Nvidia HBM Demand

Samsung Electronics plans to use a second factory in Taylor, Texas, to make components for high-bandwidth memory used with AMD and Nvidia AI processors, according to South Korea's Maeil Business Newspaper. Reuters reported that the proposed plant has orders for all planned output, with most of the volume tied to AMD. Samsung declined to comment.

The qualification matters because the report is not a public contract announcement from Samsung, AMD or Nvidia. It describes industry-source information about a facility that is years from production. Investors should treat the customer names, output commitments and schedule as reported plans rather than confirmed revenue.

Construction is expected to begin formally in November 2026. The facility would cover about 2.7 million square feet, roughly 1.5 times the size of Samsung's first Taylor plant. Equipment installation is planned from February 2029, with mass production targeted for 2030. Maeil said the factory would use Samsung's advanced 2-nanometer process.

The reported product is also narrower than a new supply of complete AI chips. Samsung would manufacture base dies for high-bandwidth memory packages. A base die sits beneath stacked memory layers and helps manage data movement between the memory and the processor. AMD and Nvidia would still rely on other suppliers and manufacturing stages for the full accelerator system.

If the report proves accurate, the plant would give Samsung an important U.S. role in the AI memory supply chain and could provide AMD and Nvidia with another domestic source for a critical HBM component. The claim that the entire planned output is already allocated would also reduce utilization risk for a very large facility. Neither the annual capacity nor the value and duration of the reported orders was disclosed.

Samsung's first Taylor factory has a separate $16.5 billion agreement with Tesla to manufacture next-generation AI chips. That contract shows the site can attract a major customer, but it should not be confused with the unconfirmed AMD and Nvidia plans for the second facility.

The schedule limits the near-term financial effect. Mass production in 2030 means the project cannot relieve current HBM constraints or contribute meaningfully to AMD or Nvidia supply in the next several years. Construction execution, equipment qualification, yields and customer validation all remain ahead.

For investors, the report is best read as a long-range demand signal. It suggests that major AI-chip designers are planning for sustained HBM requirements and that Samsung is willing to add U.S. capacity to compete for that demand. The investment case will strengthen only when the companies disclose contracts, capacity, economics and qualification milestones.