Spelling suggestions: "subject:"fabless"" "subject:"fabricless""
1 |
半導體晶圓廠投資策略與預測莊坤榮 Unknown Date (has links)
在全球持續電子化的過程中,台灣一路扮演著落實有效製造與實現設計的推手,無論從主機板、被動元件、面板、晶片設計、晶圓製造以及封裝甚至高精密組裝無所不包,在如此資本密集產業,如何操作才能達到供需平衡,為整體產業經濟創造出良性的競爭平台,避免惡性競爭,就成了不可輕忽的課題。
本文以晶圓產業為探討對象。全球產能平均利用率多年來總維持於中檔 (88%), 且平均銷售單價只能緩降而無強勢反彈,近年企業無不減少資本支出來度過低潮,整合元件製造廠(IDM)相繼喊出工廠資產輕化(fab-lite)的營運策略,這時我們的命題即是:要不要繼續投資?如何調整價值鏈?
本研究中,我們會先檢視目前市場對半導體成長預測的準確度,再經由產業價值活動代表性指標回歸分析法對相關參數做一整合之解析,對晶圓需求量與銷售價建立配適之模型,找出先行指標來達到預測,並定義上下限以供快速比對分析,最後再根據分析結果提出可能之產業趨勢議題。 / This thesis analyzes the semiconductor industry growth worldwide. The leading index via regressions has been established to achieve a reliable forecast on worldwide ASP, wafer demand and revenue.
In the long run, we expect the semiconductor demand will continue to grow at CAGR 8% (compound annual growth rate) and display less extreme cycles than past decade. However, revenue’s CAGR might be diluted to around 5%, lower than demand’s growth. Moreover, it might go down to zero-growth for some times since the ASP still slightly trend down before emerging market demand really expanded.
Continuous outsourcing is one possible solution for IDM to be fab-lite, since fab’s fix charge is billion- base that needs high utilization to maintain break-even operation. But what is the solution for foundry side to avoid ASP erosion all the way down? Our analysis identifies a need for executive managers to well predict the demand on capital expenditure when making decision.
|
Page generated in 0.0296 seconds