Recently, rumors have been circulating that Intel is in talks with Samsung to form a semiconductor foundry alliance aimed at countering TSMC’s market dominance. This collaboration between two industry giants, as described by TSMC founder Morris Chang, has sparked discussions across the sector. The key question many are asking is whether this alliance can succeed: “Can two struggling gorillas come together for warmth?”
In analyzing the situation from Taipei’s industrial sector, experts point out that both Samsung and Intel operate under the Integrated Device Manufacturer (IDM) business model. However, this model is not without its challenges; as one analyst noted, “1 plus 1 does not always equal 2—sometimes it can even be less than 1.” The idea of two struggling companies joining forces does not necessarily result in a winning combination. Chang himself has previously expressed skepticism about the prospects of this partnership, attributing concerns to fundamental issues with the business model.
Industry observers highlight the differences between TSMC’s pure foundry model and Intel’s IDM approach. TSMC has built its business on not competing with its customers, while Intel finds itself in competition with them as it seeks to expand its foundry services. Intel’s drive to stay competitive against rivals like AMD and NVIDIA has led it to increase outsourcing to TSMC, who is now handling Intel’s cutting-edge 3nm and, potentially, future 2nm processes.
Samsung, too, operates under the IDM model, producing its own chips. However, it has encountered issues recently with its manufacturing processes, resulting in a decline in market share for its self-produced chips.
If Samsung and Intel were to collaborate, three potential scenarios have been identified. The first involves Samsung taking on some of Intel’s product line orders; however, the yield rates of Samsung’s foundry services remain uncertain, which could adversely impact Intel’s product competitiveness.
The second scenario sees Intel handling a portion of Samsung’s orders. Yet here, too, Intel’s yield rates are unknown, which might not add value to Samsung’s offerings. In an effort to enhance its competitive edge, Samsung has even resorted to using high-end Qualcomm chips, the majority of which are manufactured by TSMC—creating a unique situation where Samsung phones feature “TSMC Inside.”
The third scenario involves mutual licensing of technology and patents between the two companies. This aspect warrants attention, as any potential patent disputes could negatively impact the industry. Particularly, Intel holds significant patent value in certain advanced and mature processes, which may give it an advantage in any intellectual property conflicts.
Despite Samsung’s considerably smaller share in the foundry market compared to TSMC, it does possess advanced manufacturing capabilities, such as its 3nm Gate-All-Around (GAA) technology. Analysts argue that Intel, starting from a “zero market share” position, faces the challenge of negotiating with Samsung, which is no easy task. Intel might need to offer substantial incentives to make any partnership attractive.
Intel’s most significant bargaining chip lies in its substantial outsourcing needs for its own chips. Should Intel choose to divert orders to Samsung in order to curry favor, this could lead to a reduction in TSMC’s share of Intel’s outsourced orders.
Moreover, the relationship between TSMC and Intel has long been complex. Chang has often remarked on the irony of Intel, once the industry leader, now seeking foundry services—particularly as Intel previously dismissed foundry operations as a lesser path to growth.
The founding of TSMC is also intertwined with Intel’s earlier development. During TSMC’s fundraising efforts in 1985, Chang made overtures to Gordon Moore, one of Intel’s co-founders, seeking investment, but was ultimately turned down.
