According to the KeyBanc Capital Markets report, Telco's N2 process currently shows outstanding leading advantages. As of mid-2025, Telco's N2 process yield reached about 65%, which significantly surpassed its main competitors, allowing Telco to continue to maintain its leading position in the global wafer foundry field.
The report pointed out that in order to consolidate and expand this leading advantage, NTU is continuing to invest in improving the yield of the N2 process. The company's goal is to further increase the yield of the N2 process to nearly 75% by 2026. To achieve this goal, TEK's engineering engineers are solving multiple technical challenges, including stitching problems and overlay control in multi-pattern ultraviolet light (EUV) exposure. These technical improvements are important to ensure the accuracy of the circuit pattern on the wafer, and directly affect the yield and performance of the final product.
In addition, TEK is also performing comprehensive optimization of tool and process levels, such as deploying advanced films to reduce the impact of mask particle contamination. These detailed optimization measures jointly promote the yield of N2 processes to a higher point, ensuring that tyrene maintains strong competitiveness in the future chip manufacturing market.
Compared with Telco, Intel's Intel 18A process yield is currently ranked second, and its progress is also impressive. As of the time of the report, the yield rate of the Intel 18A process was 55%. It is worth noting that this number reflects its goal of significant improvement in the previous quarter's 50% yield base, which represents the positive results of Intel in process optimization and defect reduction.
Intel's future development route includes a series of ambitious plans. Among them, the Panther Lake processor is expected to start mass production using the Intel 18A process by the end of 2025. Report expectations that Intel has the potential to push the yield of Intel 18A process to 65% to 75% through continuous process optimization and defect reduction measures. If this change can be realized, Intel's yield will surpass Samsung.
Looking forward to a more remote future, Intel has planned to launch an improved version of Intel 18A-P in the second half of 2026, which will be tailor-made for a large number of crystal foundry customers. Intel 18A-P's technical enhancement will involve modified masks and further smoothing of the flap edge, which aims to reduce pattern errors and defects. If the Intel 18A-P process can maintain a comparable yield to the original Intel 18A process, the market's expectations for Intel's round-the-loop foundry business will significantly improve.
The report emphasized that although there are reports in the industry that Intel may accelerate its development route and jump to the Intel 18A-P and enter the Intel 14A process directly, the report believes that this situation is unlikely. The report pointed out that the competitiveness of Intel 18A-P against the TECHN2 process, as well as the time when the Intel 14A process is expected to start production at the end of 2027 or early 2028, make the strategy of skipping the stage both risky and complex operation. This means that Intel is moving more vigorously into its established route and gradually improving its technological maturity and yield.
Finally, in a clear contrast to Telco and Intel, Samsung's 2-nanometer process. The process technology with its SF2 nominated performance has shown an intentional yield increase. Currently, SF2 production is maintained at about 40%, which is far below the NTF and Intel level. Reports that the poor yield performance of Samsung SF2 is due to multiple factors, including round-level defect problems, and a slow improvement in EUV patterning capabilities.
Samsung's next-generation 2nm stage process is currently expected to be launched in early 2027. However, in order to maintain its competitiveness in the advanced process field, Samsung will need to achieve a quality yield improvement before that, which is a heavy task for Samsung. Because it not only needs to solve the current technical bottleneck, it also needs to accelerate the EUV-related capacity construction and yield optimization, so that it has the opportunity to move on to the tyre and Intel's pace.
In summary, as of mid-2025, Telco's N2 process was far ahead with a yield of 65%, and set a higher yield target. Intel's Intel 18A process is rapidly catching up with a yield of 55%, and its route shows potential to achieve similar levels of NTD in the future. As for Samsung's SF2 production process faces a serious yield challenge, only 40%, and its future competitiveness will depend on whether it can achieve significant technical breakthroughs and yield improvements in the short term. This advanced process yield competition is still underway, and its results will have a profound impact on the pattern and future development of global science and technology industries.