Samsung Texas fab seen as key alternative as artificial intelligence chip demand strains TSMC

Deutsche Bank analysts say sustained artificial intelligence hardware demand is stretching TSMC’s advanced capacity and pushing major fabless chipmakers to consider Samsung’s Texas fab as a primary alternative. The bank expects TSMC’s share of leading-edge foundry work to slip slightly as customers seek additional 3 nm and upcoming 2 nm and 4 nm capacity.

Deutsche Bank analysts are warning of a persistent supply-demand imbalance in the semiconductor market, driven in large part by continued demand for artificial intelligence hardware. According to the summarized findings, TSMC’s ‘share of the advanced-node foundry market’ could fall from an impressive 95% level, down to a mere 90% in the coming years, reflecting pressure on its ability to keep up with surging orders rather than an erosion of its technological lead. The report notes that the industry-leading foundry still holds a prime position, but its global manufacturing footprint is struggling to match requests from fabless clients.

The bank believes that the Taiwanese foundry has an order backlog extending way into 2027, with many customers specifically seeking 3 nm wafer products. TSMC reportedly aims to get 3 nm node production lines pumping out around 190,000 wafers per month, by the end of this year, but the analysis suggests demand will still outweigh supply even if this goal is achieved. While earlier artificial intelligence chip production bottlenecks related to CoWoS packaging have largely been addressed, observers cited in the report argue that high-priority customers are now considering alternative foundry partners in order to secure capacity and diversify supply.

Within this context, Deutsche Bank points to Samsung and Intel foundries as the next best options, despite reports of early-stage challenges at both companies. The analysts identify a ‘big six’ group of key fabless customers, namely AMD, Apple, Broadcom, MediaTek, NVIDIA, and Qualcomm, which could be actively exploring additional ‘production capacity’ paths. Samsung’s 2 nm (SF2) and 4 nm node process lines are presented as particularly attractive, even though, like TSMC, the company concentrates its most advanced sites and expertise in its home region. A notable prediction in the report is that ‘for customers seeking alternative supply, Samsung’s Taylor fab is more likely to become the first destination’, with the expectation that the big six will prioritize Samsung’s Texas-based foundry over Intel’s best domestic site, which is described as potentially focused on 18A.

55

Impact Score

Flexible data centers could ease grid bottlenecks

Startups, utilities and chipmakers are testing ways for computing facilities to reduce electricity use during grid stress. The approach could speed connections, but critics warn it cannot replace new generation and transmission.

AMD and Rackspace plan dedicated AI compute rollout

AMD and Rackspace have finalized a phased deployment for dedicated AMD-based compute across Rackspace data centers. The capacity is aimed at regulated enterprise workloads, including clinical AI and large-scale inference.

Lexar tests SSD offloading for local AI models

Lexar is developing an AI-focused SSD approach designed to cut DRAM demand when running large language models on consumer PCs. Internal tests show the company’s storage offloading can load models that traditional local frameworks struggle to run with limited memory.

Contact Us

Got questions? Use the form to contact us.

Contact Form

Clicking next sends a verification code to your email. After verifying, you can enter your message.