半導體晶圓代工廠商微影製程開發策略之分析： T公司個案研究

Abstract

專業晶圓代工公司除了同業競爭之外，還要面對IDM公司的跨業競爭，而維持競爭力的關鍵之一就是微影技術的開發能力，在傳統光學已經接近物理極限的此時，次世代微影製程發展策略關係公司未來，如果線路微縮能力中止，摩爾定律也會隨結束，如何在眾多的次世代技術與不同供應商中做選擇，決策者要對先進技術有深入的瞭解與使用正確的評估過程。 微影技術開發中最關鍵就是曝光設備，作者以主要供應商做研究，從市佔率變化及新技術發展能力去了解競爭力之起落，再利用BCG矩陣來分析其各個事業未來之發展走向，以為決策之參考,其中，最大設備供應商ASML公司在過去的十多年內，不斷以併購與入股方法去創造其產品價值，提高其價值鏈。 爾後，Nikon與ASML間的專利戰，分析其目的發現對台積電的影響不大， 台積電在先進製程的主要競爭者分析結果，先進製程的主要競爭對手並不是專業晶圓代工廠商，反而是跨界搶市場的先進IDM公司三星和英特爾公司，作者利用成本分析法去評估兩者的競爭力，發現三星才是台積電最大的競爭對手，英特爾的高成本與客戶競合關係會影響其競爭力。三星對次世代EUV微影技術展現強大的企圖心，於其7nm先進製程就會投入使用，相較於三星的策略，台積電基於現有技術而採用漸近式策略，勝算會比較大。ASML提出的客戶共同投資合作計劃，邀請英特爾、台積電與三星入股以分散風險，讓ASML也發展出獨家的EUV技術，而台積電等公司得到高額的投資回報，可謂是雙贏的策略，業界的共同合作成為一個可行的發展途徑。 如同台積電推出浸潤式微影技術而完全改寫微影技術發展，甚至是半導體業界的路線圖，對於次世代的發展，應該鼓勵破壞式創新，以創造更有價值的技術與專利來強化競爭力，B計劃是有必要的，台積電在先進微影製程開發上已經沒有參考學習的對象，面對新製程開發的未知數，錯誤是難免的，培養在短時間內修正方向的應變力，可以提高企業存活率。競爭對手成立專業的投資部門或是投資公司是一個好的策略，台積電應該要有專責單位，去周延的調查與評估半導體相關的新技術，與合適的投資機會，以期在早期就可以掌握新技術，也可以去扶植第二設備供應商，以避免設備供應商出問題或是出現價格問題。

This research proposed long term strategy for TSMC keeping its competitive advantage on lithography technology. TSMC competes not only pure foundry manufacturers, but also global IDM companies in this area. How to choose right exposure scanner supplier to develop next generation process technology among dynamic competition, this researcher adopted appropriate business models with strong technology background to propose the best solution for TSMC. First of all, this researcher evaluated market share change and innovation ability among exposure suppliers with consideration of their related business development. Then, the biggest supplier ASML was defined the best partner as its ability to create value chain by merge and accusation of other technology companies. Secondly, this research pointed out that the key threaten for TSMC was not other pure foundry manufactures, but leading IDM companies, such as Samsung and Intel who includes integration, design and manufacturing ability of foundry. By analysis production cost and deliverable products, Samsung was defined the biggest competitor for TSMC while Intel played in different battle filed on premium tier products with high cost. Intel was also the competitor of other foundry buyers which also limited its development on its market share on foundry manufacturing market. The research revealed Samsung showed strong ambition on next lithography technology development with skipping different generation foundry size production, so Samsung might close up the current technology disadvantage with TSMC. The author believe TSMC chose right approach on steadily narrowing the foundry size production. Thirdly, ASML created a new win-win model by inviting its core clients, including TSMC, Samsung and Intel to core development on next generation foundry scanner. This new customer co-investment program split high R&D cost and create new EUV technology which strengthen ASML’s leading position among another foundry scanner manufactures. ASML’s clients, such as TSMC, also benefits from this program to enjoy better scanners and gain money return via investing ASML’s co investment program. Finally, this research encouraged TSMC to continue disruptive innovation to sustain its technology leading position as TSMC already demonstrated this ability on setting up the “roadmap” among semi-conductor industry. Moreover, TSMC should enhance it’s dynamic change ability as technology change fast and consider to develop 2nd exposure scanner to avoid it’s high dependency on ASML. Continuing monitor the industry changes and early to involve new technology with right suppliers might sustain TSMC’s competitive advantage in pure foundry provider market.