賠錢都得做——台灣積體電路製造公司

:::

1987 / 6月

文‧張毅君 圖‧邱勝旺


俗語說:「賠錢的生意沒人做。」

 

有一家公司,光是投資額就高達一百億新台幣,卻有人認為「賠錢都得做!」


這就奇了!

今年二月,有一家「新店」開幕了。

這個店來頭可不小,雖然是「寄人籬下」,而且門面不怎樣,但它的投資額可是十分驚人。

這店的名號叫「台灣積體電路製造公司」(又稱TSMC),暫時棲身在工研院電子所裡,主要的工作就是製造積體電路。

目前每月生產六吋晶圓(積體電路的電製造須在矽晶圓上完成,再切成晶片)七、八百片,產量不大,但預計年底可達一萬片。

TSMC還有第二家店,目前在新竹科學園區興建中,預計七十八年底完工。它也不是什麼「省油的燈」,投資額同樣高得令人瞠目。

算盤打一打,不得了,這二家投資額合計約一百億新台幣。這是什麼樣的大事業,值得花這麼多錢?

資訊工業之「米」

如果說這是我國資訊工業命脈所系,再大的投資恐怕也在所不惜。

廿一世紀是資訊工業的世紀。行政院政務委員李國鼎先生,在今年五月召開的「一九八七年國際超大型積體電路技術系統暨應用研討會」開幕致詞中指出,自一九七九年至一九八六年,全世界電子設備的年收入由大約一千億美元,迅速成長到超過二千億美元;而在許多國家中,電子工業所雇用的人員數目,已遠超過其它工業所雇用的人員。

中華民國政府也早已將資訊工業訂為策略性工業,並設有工業技術研究院電子所及資訊工業策進會,在我國資訊工業的發展上扮演主導的角色。去年,我國電子產品總產值達六十九億美元,居各類產品的第一位。

而積體電路(IC),尤其是功能複雜的超大型積體電路(VLSI)則是電子產品的基本元件。

過去,國內的IC工業以下游的裝配業最繁榮。大部分業者都向國外買進IC,組成電路板後再出售。

在一九八三年的一次全球IC缺貨的事件中,我國業者即因無法取得足量的IC,直接影響了電子資訊產品的成長。縱使廠家手中握著一堆訂單,卻獨缺最主要的原料——IC,這樣的處境,實與巧婦「等米下鍋」時的心情無異!

之所以被人牽著鼻子走,最大的關鍵在於我們無法大量製造IC。

其實,中游的IC製造業,我國並非沒有,但只有聯華電子和工研院電子所生產,數量並不多,只能應付國內電子資訊業百分之廿的需要;而且技術層次也不算高,大部分是做電子錶、電子玩具、音樂卡等消費性IC。談到電腦內部所需的主要元件——記憶晶片,我國在設備上、技術上都不夠。

至於上游的設計開發,情況也不佳。

沒有大規模的製造工廠?

三年前,在美國矽谷的一批海外學人,帶著他們設計的超大型積體電路回國,成立了「華智」、「國善」和「茂矽」三家超大型積體電路公司。

但是,由於他們未如預期籌措到建廠資金,所設計開發的新產品無法自己製造;而當時國內又沒有大規模的製造工廠。於是,他們委託日、韓工廠生產,結果大筆的鈔票被賺走了。

「沒辦法呀,台灣缺少大量生產的設備,我們產品的生命週期又很短,必須馬上生產,才有可能回收」,華智總經理何峻解釋。

為什麼我們沒有大規模的製造工廠呢?

「因為,IC工業是資本密集、風險性高的行業」,茂矽電子的行銷與市場部經理沈文義說。

風險性高,是因為高度競爭。由於產品技術一日千里,機器設備的投資動輒上百億,為了保持市場佔有率,必須把大把的銀子往塈諢C如果投資稍慢,新的技術、產品開發出來,舊產品被打入冷宮,公司就很難生存了。

但IC工業卻也是很賺錢的行業,它的附加價值超過百分之百,比平均百分之廿五的一般製造業高得多。

但是,天文數字般的投資額,一般公司誰敢輕易嘗試?尤其是在以中小型企業為主的我國。

「政府不插手不行啦!」許多業者反應。

日本就是一個好例子。今天它在世界的半導體市場上有百分之七十的佔有率,政府的功勞不小,在它發展超大型積體電路的起步階段,政府投資額佔了一半。

需要能聯手和日本作戰的朋友

委託日、韓生產,錢被賺走的這一條「導火線」,引發了政府發展超大型積體電路的決心。這項由前經濟部長徐立德提議投資的計畫,就這樣「動」了起來。

民國七十四年九月,正式宣佈將成立一家超大型積體電路公司,並由工研院院長張忠謀負責籌劃。

公司採民營方式,為世界上的超大型積體電路設計公司「代工」生產。政府投資百分之四十八.三。

另外一半的股份,荷蘭飛利浦公司佔百分之廿七.五,國內幾家民營公司佔百分之廿四.二。

「為什麼是飛利浦?」有人問。

飛利浦的消費電子產品在歐洲一直保持最高市場佔有率,也是唯一能和日本廠商比高下的歐洲公司。

目前,飛利浦跟西門子在歐洲合作,研究發展百萬位元的超大型積體電路,精密程度已經小於一微米了。

飛利浦總裁范戴克曾提及在台投資的原因:「我們需要能聯手跟日本作戰的朋友,中華民國就是這個理想的伙伴」,他分析,因為我政府很支持這個廠,而且台灣技術人才的水準整齊,聰明才智和教育水準都很高。

「我們覺得這樣一個有成長慾望,又有人才的地方,很值得投資」,他說。

合作開發是未來的趨勢

張忠謀表示,飛利浦公司的「入股」,我們可獲得他們技術的轉移;另方面由於研究開發的成本愈來愈高,這種合作開發的模式已經是世界半導體工業發展的趨勢。而飛利浦公司所扮演的角色,除了是投資者、技術伙伴(提供二微米以下的技術轉移),還是客戶,每年將委託TSMC製造一定數量的超大型積體電路。

在張忠謀的構想中,TSMC定位在「製造服務」業。它只接受客戶委託「加工」,不具有自己的品牌、產品,也不作產品設計和銷售。這樣就避開和日、韓等「一貫作業」的IC公司正面交鋒。

就像做衣服的裁縫,客戶訂做什麼就做什麼,至於是鳳仙裝,還是洋裝都沒關系。

這樣的經營方向,風險小;目前市場雖沒「成衣」大,但市場正在成長。

根據估計,到一九九○年代,訂製型積體電路的需求將大增。

紐約時報也曾報導,在全球電子業不景氣時,目前美國加州矽谷少數幾家仍賺錢的超大型積體電路廠商,都是接受客戶訂製的工廠。

TSMC掌握了先機

有人擔心:TSMC志在製造,但「開市」以後,到那堨h拉那麼多筆生意呢?

有一批人,是TSMC最可能的客戶。

他們以往找一貫作業的工廠製造IC,但產品卻被對方抄襲。像美、日在超大型積體電路工業發展多年,就常發生這種因抄襲而引發的訴訟官司。

「TSMC的特色是不設計,不開發,可消減客戶對產品被抄襲的顧慮」,目前兼任TSMC董事長的張忠謀說。如此一來,「安全度」就成為我們極有利的條件。其次,由於半導體工業投資額太大,許多小公司無力建造自己的工廠,就可以將開發出來的產品,委託TSMC製造。「在半導體工業分工愈來愈細的趨勢下,TSMC無疑地掌握了很好的先機。」張忠謀認為。

而飛利浦台灣分公司總裁貝賀斐表示,TSMC如果能提供保護智慧財產的保證,就能爭取到更多客戶的信任。

除了這些,我們手上還有那些籌碼呢?

台灣有什麼競爭條件?

在半導體的成本競爭上,良率(成品品質優良的比例)往往是勝負的關鍵。因為在電路纖細到微米(萬分之一公分)為計算單位時,我們肉眼看不到的灰塵粒子,對IC來說,卻彷彿一顆大石頭壓在電路上,而那片IC就算報銷了。

在積體電路公司來說,良率若能達到百分之五十,就算是相當不錯的表現。正因為如此,在競爭上,誰的良率高,成本自然就可以壓低,取得優勢。

機械、設備的精良固然是提高良率的重要因素,但工作人員的素質更是重要。

美國在半導體戰場上由優勢轉為劣勢,輸給日本,正是因日本工人的團隊精神、專業水準、流動率低是美國所望塵莫及的;也難怪有人會說「一九七○年代,科技中心在矽谷,但九○年代,科技中心將移到日本。」

TSMC總經理戴克斯對這點表示了他的看法。就良率來說,台灣的工程人員素質、守紀律的精神,不比日、韓差,所缺的僅是經驗與生產技術的穩定度。目前試尷熊痕G,良率都還不錯,他相信假以時日,一定還會有更好的成績。

華智總經理何峻表示,目前他們也負責TSMC的一部份測試工作,良率還不錯,所以他在考慮將開發出來的產品,委託TSMC製造。

年輕的工業奇兵

但是,也有人懷疑,TSMC接受客戶各式各樣的產品製造,本身並不涉及產品開發,技術層次要如何提昇呢?

「要一步一步來」,張忠謀強調,TSMc並不是我們半導體工業的「終點站」。

「因為,沒有一個經濟發達的國家是光做製造的」,張忠謀指出,要先鞏固製造生產方面的基礎,再分高科技、自動化、高品質三方面進行,讓我們的技術升級。除了製造生產外,還要做產品的行銷。等腳步站穩了再來談成立「一貫作業」的公司,開發生產積體電路。

而今天TSMC在製造上,除了自己本身的製程技術外,一方面也可以吸收委託製造廠商轉移的技術。例如在製造尖端技術的記憶晶片時,TSMC的工程技術人員同時也可以得到設計開發方面的經驗,如果未來有可能成立「一貫作業」的公司,這就是很好的奠基工作。

所以,張忠謀認為,「代工」聽起來層次很低,但在超大型積體電路工業來說,卻可以由一些經驗、技術的累積,儲備優秀的人才。而成立「一貫作業」的公司,正是我們主要的目標。

「有了一貫作業的公司,才能生產自己的產品,開創自己的品牌,才能在國際市場和他人競爭。」張忠謀表示。

有人形容TSMC是一支年輕的「工業奇兵」,雖然經驗不夠豐富,但卻有無窮的潛力和衝勁。

年輕,是它最大的本錢!

〔圖片說明〕

P.68

上圖中工程師手上握的即是一片矽晶圓,可切割成二百多塊如下圖中食指上的積體電路(IC)。設計師後面那張複雜的電路設計圖,濃縮在那片小小的IC上,是電子資訊產品中不可或缺的零件。

P.69

工研院院長張忠謀是此次籌建百億大廠的「靈魂人物」。

P.71

精密的設備與技術,是超大型積體電路工業所必備的條件。

P.72

工作室的潔淨度,是影響產品良率的重要因素之一。

P.72

TSMC總經理戴克斯先生,在半導體界有其獨特的經營長才。站在他身後的是兩位副總經理,左為曾繁城,右為洪鐵成。

相關文章

近期文章

EN

The Foundation for Taiwan's IC Manufacturing Industry

Jack Chang /photos courtesy of Chiu Sheng-wang /tr. by Phil Newell


This February, a "new shop" opened in Taiwan. Although the storefront wasn't anything special, the investment behind it was stunning. This new company is called the Taiwan Semiconductor Manufacturing Company (or TSMC).

The company is temporarily housed in the Electronics Research and Service Organization of the Industrial Technology Research Institute, and its main work is the production of integrated circuits. At present it produces about seven or eight hundred six-inch silicon wafers a month. (In producing circuits, it is necessary to first complete them on the silicon wafers before they are cut into chips.) By the end of the year, it is estimated that production could reach 10,000 total units.

TSMC is building a second "shop," in the Hsinchu Science-Based Industrial Park. The combined investment for these two establishments is approximately NT$10 billion (about US$300 million). The reason for such an enormous investment is simple: without it, there is a danger of falling behind.

Minister Without Portfolio Li Kuoting this May hosted a special conference on "Very Large Scale Integration [VLSI] Technology Systems and Their Uses." In his opening statement he noted that between 1979 and 1986, the global income in the area of electronic equipment went from US$100 billion to over twice that amount. Moreover, in many countries employment in the electronics industry far exceeds that of other industries. And it is likely that the industry will only increase in importance.

Now the government of the ROC has already decided to focus on electronic information systems as a key industry. In order to stay on the "cutting edge" of this technology, it is necessary to get into the VLSI field. And that is why it is so important to establish this new company.

In the past, Taiwan has been primarily doing downstream assembly of circuit boards from imported IC's. The boards are then reexported. But this industry is full of difficulties, as was revealed during the 1983 worldwide shortage of IC's when many assembly operations were left without necessary parts. At the midstream stage, Taiwan itself produces few IC's-- far less than domestic demand--and most of these end up in low-tech items such as toys and wristwatches. As for the upstream stage of research and development, the current situation is also not satisfactory.

Three years ago a group of overseas companies such as Vitalec, Quasel, and Mosel wanted to return to Taiwan to establish a facility for producing VLSI IC's. Since they were unable to raise the necessary capital to produce the items themselves, they needed help. Unfortunately, at that time Taiwan lacked the large-scale production facilities required, and the VLSI's ended up being produced in Japan and Korea. As Wayne Shen, manager of sales and marketing at Mosel, pointed out, it was hard to find private investors because the IC industry requires a large, high risk investment.

This occurrence spurred the government to make a decision to develop a local VLSI industry. The decision to establish a VLSI company was formally announced in September of 1985, and the head of the Institute for Industrial Research, Morris Chang, was assigned responsibility for planning.

It was decided that the company was to be publicly run, and its main job to undertake production for VLSI design companies. The government share of the investment was 48.3 percent. Philips Company of the Netherlands put up 27.5 percent, while various private companies in Taiwan provided 24.2 percent. Philips makes for a logical investment partner because Philips consumer electronics products already dominate the European market, and it is the only company in the world which can really compete with the Japanese in this field. From the point of view of the president of Philips, Cor vander Klugt, Taiwan is a logical partner both because of the strong government support for the project and because of the high standards and skills of Taiwan's labor force.

Besides being an investment partner, Philips is also a channel for technology transfer to Taiwan. There is a trend toward this style of cooperative venture in the semiconductor industry because of the extremely large financial base for R&D which can be established.

TSMC is in the "manufacturing service" industry. That is, it does the manufacturing work on special order from design companies. It has no particular brand name or product, and does neither design nor marketing. In this way it is quite different from the integrated operations of Japan and Korea.

The potential market for VLSI is quite large. Likely customers for the services of TSMC would include the large number of Japanese and American companies who find that integrated operations infringe on their designs and steal their R&D. In this sense, notes Morris Chang, who is concurrently chairman of the board of directors of TSMC, this high degree of product security may be TSMC's greatest lure. Furthermore, many design firms cannot afford their own manufacturing operations. TSMC can fill this gap.

In addition, in the crucial area of efficiency (determined in the VLSI industry by the percentage of units out of the total produced of which the quality is satisfactorily high) preliminary indications are that Taiwan's capability is quite good.

Morris Chang notes that manufacturing is only a first step. He hopes that Taiwan can progress step by step from manufacturing through automation, improved quality, and higher technology at first, and then sales. TSMC's technical and engineering personnel will also benefit from exposure to transferred technology, and can gain some experience in questions of design. After these steps have been completed, one can begin to speak of setting up "integrated operations." Only then can Taiwan reach its main goal of producing its own products under its own label to compete in the international market.

Though TSMC is still a bit short on experience, its potential is boundless. Indeed, youth is its greatest asset.

[Picture Caption]

This engineer is holding one silicon wafer, which can be cut into over 200 IC's like the one below. What you see behind the engineer is a complex circuit design map, which is an enlarged version of what is on a chip.

The director of the Industrial Technology Research Institute, Morris Chang, is the driving force behind the ROC's NT$10 billion IC project.

Minute equipment and technology are key features of the VLSI industry.

The cleanliness of the workshop is one of the most important factors affecting the efficiency and wastage in production.

TSMC's General Manager James Dykes has had long experience in the world of semiconductor management. Behind him are two assistant general managers: Tsung Fan-cheng (left), and Hung Tie-cheng.

 

X 使用【台灣光華雜誌】APP!
更快速更方便!