TSMC's A14 Process Advances Yield and Performance, Attracting AI and Smartphone Demand

TSMC’s A14 process node is maturing faster than its N2 predecessor, with internal test vehicles demonstrating nearly 90% device performance and SRAM yield as the foundry prepares for 2H 2028 mass production.

Chips & Climate Desk
Updated Nov 12, 2026 2 min read
TSMC's A14 Process Advances Yield and Performance, Attracting AI and Smartphone Demand

TSMC has reported significant technical progress on its 1.4nm-class A14 fabrication process, indicating that development is tracking ahead of the N2 node at a similar stage. During a recent earnings call, TSMC CEO C.C. Wei confirmed that internal product-like test vehicles have achieved nearly 90% of target device performance and close to 90% yield on 256Mb SRAM structures. This represents a notable improvement from April, when the node reached over 85% target transistor performance and over 80% SRAM yield, suggesting the process is maturing rapidly approximately 2.5 years before its scheduled mass production in 2H 2028.

The accelerated maturation of A14 compared to N2 is likely driven by TSMC’s accumulated experience with gate-all-around (GAA) nanosheet transistors. While N2 served as the foundry's initial implementation of GAA technology, A14 incorporates the second generation of these devices. This evolution allows the process to leverage refined transistor designs and manufacturing techniques established during the N2 development cycle. While high SRAM yields on test vehicles indicate improved defect density and process uniformity, they do not directly equate to the functional or parametric yields expected for commercial processors.

Technically, A14 combines these second-generation GAA nanosheets with a new standard-cell architecture. TSMC projects that this node will provide a 10% to 15% performance increase at the same power and transistor count compared to N2, or alternatively, a 25% to 30% reduction in power consumption at equivalent frequency and complexity. The process is also expected to offer a transistor density increase of approximately 20% for mixed designs and 23% for logic. Notably, A14 will not feature Super Power Rail backside power delivery.

Despite the absence of backside power delivery, the node is seeing broad adoption interest. According to Tom's Hardware, TSMC is observing strong engagement from both the smartphone and AI/HPC sectors. CEO C.C. Wei noted that customer tape-out activity is currently ahead of schedule, signaling confidence in the process roadmap. This early momentum could potentially allow TSMC to initiate high-volume manufacturing ahead of its original timeline, or provide the foundry with superior functional and parametric yields when production begins.

For engineers evaluating the foundry's roadmap, the A14 progress highlights a shift toward more versatile nodes capable of addressing diverse workloads. By balancing high transistor density with power efficiency, TSMC aims to support both mobile SoCs and high-performance AI accelerators. As the industry continues to navigate the complexities of GAA adoption and node scaling, the rapid maturation of A14 suggests that TSMC is successfully leveraging its N2 manufacturing experience to shorten the development cycle for its next-generation process technology.

Sources

  1. 01 TSMC confirms significant yield and performance improvements in A14 update — strong interest from AI/HPC and smartphone customers — Tom's Hardware