台灣半導體設計服務業之競合發展策略研究

Abstract

IC產業發展至今已近半世紀，產業型態由原先的垂直整合模式朝向專業化分工演進。80年代初出現無晶圓廠的IC設計公司專司晶片設計工作，末期則出現晶圓專業代工廠，產業逐漸確立了晶片設計與製造分工的結構。進入2000年後，隨著摩爾定律的持續有效，製程的精進使得一個IC元件能夠容納的電晶體急速增加，設計端卻相對無法跟上如此複雜的設計，為縮短產品開發時間、提高設計生產力，一種專司提供設計業者事先經定義、驗證、且可重複使用的IP功能元件的矽智財產業於焉出現，半導體產業因此呈現進一步的專 業分工。 雖然台灣已是全球第二大半導體產業國，但相較於全球競爭者，台灣半導體設計服務業者大多仍是獲利表現不佳。本研究的目的，希望能針對IC產業的專業分工結構，透過競合理論與交易成本、垂直整合等相關理論，探討設計服務業者在新的產業分工鏈下所扮演的角色，及其與價值網攸關廠商間的競合關係，從而對設計服務業者提出將來的發展空間與方向的建議 經過深度訪談與相關次級資料蒐集之質化研究，本研究導出下述結論：(1)現行設計服務業者與其周邊廠商間的關係仍是合作多於競爭；(2)前段設計服務產業在技術上的進入障礙不大，相對的，退出障礙亦不高。後段設計服務領域，在產業的進入上則是以關係導向(與晶圓廠)為主；(3)邊際投入成本不高下，IC設計公司與IDM廠皆會是設計服務領域的潛在進入者，使得該產業領域競逐廠商之類型與數目眾多；(4)在服務幾不具差異化與附加價值不大之情況下，後段設計服務的價格競爭甚為激烈；導致業者生存不易而不得不往附加價值較大的前段設計服務領域發展；(5) 設計服務業者之上游廠商皆為其互補業者；然其中的SIP供應商與設計服務業者間又具備一定程度的替代關係；(6) 設計服務業者的上游供應商EDA業者本身亦因策略性因素而有投入前段設計服務領域，將來如產生業務上衝突，對設計服務業者而言將較為不利；(7) 對以提供統包服務(Turnkey Service)為主的業者而言，與上游晶圓廠間的關係攸關重要，並直接影響其競爭力；(8) 基於晶圓廠仍以COT客戶為優先服務對象，晶圓廠產能吃緊會不利後段設計服務業者的發展。且對於產能利用率高的晶圓廠而言，設計服務業者的存在價值對其相對較低。 最後，根據上述的研究分析結論，本研究針對台灣設計服務業者提出了下列六點策略建議：(1) 以提供Turnkey服務為主的設計服務業者應強化與晶圓代工廠間的關係；(2) 業者應往國外市場發展，以培養接先進製程的能力；(3) 進入與加強前段設計服務能力是不得不然的趨勢；(4) 一家晶圓廠體系下不宜太多家設計服務公司；(5) 規模太小下業者彼此整合是好事；(6) 業者應建立中立的服務地位，並持續拓展其IP授權業務。

Since the invention of the first integrated circuit (IC) in 1959, the semiconductor industry has been evolved from a vertically-integrated industry to horizontally configured industries based on increasing levels of specialization. In the early 1980s, specialized IC design companies appeared. But it is not until 1990, when the specialized semiconductor manufacturing company, so called 'foundry', came into play, did the fabless IC design companies prosper and pump up the industry growth. As Moore’s law continuously guides the development of semiconductor manufacturing technologies, a chip can contain more and more electrons inside and the IC design becomes even more complicated than ever. Therefore, a new design approach based on IP reuse becomes more competitive as the design productivity increases, which in turn induces the emergence of IP providers and design service companies. The semiconductor industry had literally entered its 3rd evolution of the horizontal specialization. Although Taiwan has been the second largest semiconductor industry in the world, most of its design service companies are still underperformed, compared to its global competitors. The objective of the present study is to explore the structural reasons underlying the industry development by exploring the positioning of the design service companies in the semiconductor industry’s new industry landscape as well as the competitive dynamics among value-net players. Based on field interviews and qualitative research on the broadly defined semiconductor industries, we reached several conclusions: (1) The relationship of the design service company with its business-related companies remains more cooperative than competitive; (2) The entry and exit barriers of the front-end design service are not high, while the foundry’s attitude determines major entry barrier of the back-end design services; (3) Due to the very low marginal costs, the fabless and the IDM with the IDM-foundry business model are both the potential competitors of the design service company; (4)As the result of the nearly indifferential and low value-added characteristic of the back-end design service, the price competition in this area is dramatically high, which makes the back-end design service firms earn too little and declining margin and forced them to stretch to the higher value-added front-end design service business; (5) Almost all the upper-streamed companies of the professional design service company are its complimentary partners. However, the SIP vender still has certain degree of substitution with the design service companies; (6) Two of the three major EDA venders, the DS companies’ upper-stream providers, also entered the business area of the design service business as a result of their strategic choices in the global market, and that may brought potential threat to the professional design service companies; (7) For the back-end design service company, its relationship with the foundry is vital to its competitive advantage; (8) As the foundry put higher priority on its COT customers, the growth of the design service company would be constrained if the capacity usage rate of the foundry was high and tight. Beside, the value-added of the design service company would be low to the foundry that has constantly high capacity usage rate. Base on the above analyses, we propose several pieces of advice for the design service industry in Taiwan: (1) The back-end design service company should strengthen its relationship with its cooperative foundries; (2) The company should go out of Taiwan to the global market in order to develop its ability of taking the high-end design service orders; (3) It is necessary for the back-end design service company to enter the front-end design service business area; (4) It is improper for a foundry to ally with too many design service companies; (5) For the design service firm that has too small scale and limited resources, the best strategy for it to grow and develop its capability as quick as possible is to merge with another firm that has its complimentary strength and resources; (6) To gain customers’ trust, the design service company should not have its own chip product. Thus, the IP business appears to be a good choice of diversification.