At a time when the global wafer foundry competition is becoming more and more intense, the "talent war" between enterprises It is also becoming more and more intense. Recently, Samsung Electronics hired Margaret Han, a veteran executive who has worked at TSMC for 21 years, and appointed him as the head of Samsung's North American foundry business. This move is not only an important part of Samsung's globalization strategy, but also reveals its determination to launch a new offensive in the North American foundry market. At the same time, Samsung's $17 billion advanced process fab in Taylor, Texas, will also be put into operation in 2025, aiming at TSMC's long-dominated North American high-end customer market. So, can Samsung's series of layouts really shake the North American wafer foundry pattern?
Ⅰ From TSMC to Samsung: A Strategic Shift for Senior Talent
According to LinkedIn, Margaret Han joined TSMC in 2000 and has held key positions in technology development, manufacturing integration, and customer support. She has participated in a number of core projects, including process integration and mass production introduction of 28nm, 16nm and 7nm process processes, and has played a key role in improving the mass production yield and customer satisfaction of TSMC's advanced processes.
After leaving TSMC, Han held executive positions at Intel and NXP Semiconductors, where she expanded her expertise in chip design and system integration, as well as developed a broader understanding of the North American customer ecosystem. In March 2024, Samsung Electronics officially hired her to lead the North American foundry business, with the intention of leveraging her deep customer relationship and management experience to increase her technical influence and customer penetration in the North American market.
Pictured: Margaret Han, who worked for TSMC for 21 years, is now hired by Samsung Electronics as the head of the North American foundry business
Ⅱ $17 billion investment: Samsung's North American ambitions
Han's arrival comes at a time when Samsung is expanding aggressively in North America. According to Samsung Electronics' announcement, its wafer fab in Taylor, Texas, started construction at the end of 2021 and is scheduled to be put into operation in 2025, with a total investment of up to $17 billion. This is one of Samsung's largest investments in U.S. history, and it is also an important support point for the global layout of its wafer foundry business.
The new facility will be built on the state-of-the-art GAA (Surround Gate) architecture 3nm process technology, with the goal of attracting orders from high-performance computing and cloud services customers, including NVIDIA, AMD, Tesla and Amazon. This marks Samsung's attempt to extend its foundry capabilities from the previous main battlefield of mobile chips to high-margin segments such as data centers and AI accelerators.
But in the North American market, TSMC's roots are extremely deep. According to the 2023 global foundry market data released by TrendForce, TSMC's market share reached 56.1%, much higher than Samsung's 15.5%. In addition to its technological leadership, the key to TSMC's long-term leadership in the North American market lies in its long-term stable and in-depth cooperative relationship with customers.
Figure: Samsung Electronics' fab in North America
Ⅲ Samsung's shortcomings: technical bottlenecks and trust gaps
From a technical point of view, Samsung still faces certain challenges in advanced manufacturing processes. According to a number of media and semiconductor analysis institutions such as The Information and SemiAnalysis, TSMC's 3nm process (N3B) yield has stabilized to more than 90% by the end of 2023, while Samsung's first-generation 3nm (SF3E) yield has hovered around 50% for a long time. This gap directly leads to a lack of confidence in the performance, cost, and controllability of mass production.
A typical case is the foundry choice of Google Tensor chips. Although Google initially planned to use Samsung's 3nm process on some models, after multiple rounds of trial production, due to the failure to meet expectations in yield and power consumption control, it finally chose to return the main order to TSMC. At the same time, customers such as Qualcomm and AMD have also made it clear that TSMC is preferred as a partner in advanced processes (4nm and below).
More importantly, wafer foundry customers have extremely high requirements for "trust", which not only involves advanced technology, but also includes controllable delivery, flexible production capacity, IP compatibility and customer collaborative development experience. Samsung, on the other hand, is still lagging behind TSMC in terms of customer collaboration processes, ecosystem support, and long-term project cooperation, although it is attractive in terms of price.
Ⅳ Han's role: a key variable to break the game?
In this context, can the joining of Margaret Han become a key variable for Samsung to break the game?
First of all, she established stable contacts with many major North American customers during her time at TSMC, including Apple, AMD, Qualcomm, NVIDIA, etc. This shift of trust and network in the industry is crucial for Samsung, which is looking for a breakthrough in its customers. As a technical communicator and market understander, she is able to help Samsung better understand customer needs and secure more prototype trials or low-volume foundry opportunities in the initial evaluation stage of customers.
Second, Han is well versed in the key bottlenecks in advanced manufacturing processes. She has led TSMC's optimization of production issues in the FinFET and EUV adoption phases, and her practical experience in process integration and yield improvement may help Samsung quickly accumulate technical stability in the 3nm GAA architecture. In particular, the best practices accumulated by Han in yield-critical processes such as multiple exposures, photomask calibration, and copper barrier layer stacking are expected to serve as an important reference for Samsung's improvement strategy.
However, change is not something that can be achieved by one person. The foundry project cycle is often 2-3 years, and Han's strategic role is more inclined to medium- and long-term value emergence. If Samsung wants to shake up the North American landscape, it needs to make efforts in the following aspects:
Ⅴ Samsung's future breakthrough path: the trinity of technology, customers, and ecology
1. Process stability and yield improvement are the top priorities
If it is unable to increase the 3nm yield to more than 80% by 2025, it will be difficult for Samsung to win the confidence of mainstream customers in mass production. This requires continuous investment in equipment tuning, material optimization, defect monitoring, etc., to narrow the technology gap with TSMC.
2. Customer development needs to break through from small and medium-sized customers
In view of the high threshold for large-scale customers to transfer orders, Samsung can consider starting with medium-sized AI chip start-ups or domestic alternative customers, such as Cerebras and Tenstorrent, or China's Horizon and Suiyuan, to establish a small number of cooperation first, and then gradually expand the scale.
3. Build a localized ecosystem and service team
TSMC has an extensive range of technical support centers and EDA and IP ecosystem partners in the United States. Samsung also needed to enhance the capabilities of its local customer service and support team in the United States, including a real-time response mechanism, Design Enablement platform construction, and a collaborative verification process for software and hardware.
4. Take advantage of U.S. government subsidies to strengthen policy coordination
Under the framework of the CHIPS and Science Act, Samsung has received about $6.4 billion in government subsidies (according to the April 2024 announcement of the U.S. Department of Commerce), and if it can translate the subsidy advantage into production capacity and equipment support capabilities, it will also strengthen its confidence to attract customers.
Ⅵ Conclusion: Although the battle situation is not clear, the variables have appeared
In summary, Samsung Electronics' foundry strategy in the North American market has entered a substantial stage. From the large-scale construction of factories to the poaching of senior executives of TSMC, it shows its dual ambitions of global market layout and technological breakthroughs. Although it still lags behind TSMC in terms of yield, customer trust and ecosystem improvement, the addition of Margaret Han has injected new technology genes and strategic wisdom into Samsung, which may become one of the key variables in the next two years.
But to shake TSMC's position in North America, Samsung will need to continue to invest in customer success stories, technical stability, and long-term trust accumulation. With the surge in demand in the North American foundry market driven by applications such as AI computing power, edge computing, and autonomous driving, this new battle led by technology, talent, and policy may have just begun.
