近二十年以來，手機、物聯網科技產業的蓬勃發展，推動著許多應用市場創新，因此導致數位經濟生活中所產生的巨量資料，必須運用電腦科學實證的軟體演算法來的處理及應用，加上人工智慧技術開發技術突破，以及晶片平行設計架構、運算效能等的大幅提升。使成熟的演算法進行IC產業化，人工智慧晶片也應而生。本研究的目為藉由探討各種人工智慧技術、晶片的發展模式、開發工具的現況、市場應用範圍、晶片廠商營運規模等資料，研究目前的人工智慧晶片產業，選出人工智慧晶片的技術發展、應用晶片的開發生態、晶片廠商的營運及投資策略、晶片的主要應用領域為主要的競爭影響因子。並以五家主導公司及一家新創的晶片開發公司為例，探討波特五力模型中各家晶片商在影響因素的表現或可能策略，並以創新擴散理論探討晶片的初期應用開發者是如何被創新階段的人工智慧晶片技術所說服。人工智慧晶片廠商，在技術發展上，以Intel、Nvidia以併購補強或延伸既有的晶片產品為主，強調平台型的平行晶片功能，處理問題以軟體方式達成計算，Qualcomm與Apple主要是將加速器嵌入原先手機晶片，加強晶片在處裡影像功能，Google對於人工智慧產業投資不斐，開發特定的機器學習晶片，相較其他的廠商不同之處，是以開發生態切入，新創公司Graphcore的特定推理晶片技術中，是屬佼佼者，為眾多投資者青睞。開發生態中，Google與Nvidia較為完整，Intel努力追趕中，Apple的應用軟體開發生態最為強大，如果能引導對人工智慧應用開發，能量不可小覷。總體比較，Google、Nvidia、Intel和Graphcore是較具人工智慧晶片的競爭條件。晶片的發展，目前是群雄並起的競爭態勢，除了主導廠商憑藉著既有市場優勢外，更積極研發新的晶片或著是併購新創的人工智慧晶片廠商。但許多新創的晶片公司，透過風險投資的方式，也希望在人工智慧晶片的找到應用的定位，以期能進入資金、研發的正向循環中。總之，晶片的研發成果與應用市場的交互支持，促使應用領域的迅速擴張，新創的晶片設計公司也較有機會找到應用方向，在市場中的主導廠商如何應用精準的產品策略，也是一值得探討的問題。

In the past two decades, the booming development of mobile phone and IoT technology industries has promoted many application market innovations. As a result, the huge amount of data generated in digital economic life must be processed and applied using computer science empirical software algorithms. In addition, technological breakthroughs in artificial intelligence technology development, as well as the parallel design of the wafer, computing performance and so on. The mature algorithm was industrialized in IC, and artificial intelligence chips were born. The purpose of this research is to study the current artificial intelligence chip industry and select the technological development of artificial intelligence chips by discussing various artificial intelligence technologies, development models of chips, current status of development tools, market application scope, and scale of operation of wafer manufacturers. The development ecology of the application chip, the operation and investment strategy of the chip manufacturer, and the main application areas of the chip are the main competitive influence factors. Taking five leading companies and a newly-developed wafer development company as examples to discuss the performance or possible strategies of various chip vendors in the Porter's five-force model, and to explore the initial application developers of the wafers with innovative diffusion theory. How to be convinced by the artificial intelligence chip technology in the innovation stage. Artificial intelligence chip manufacturers, in the development of technology, Intel, Nvidia to strengthen or extend the existing chip products, emphasizing the platform-type parallel chip function, processing problems to achieve calculations in software, Qualcomm and Apple are mainly accelerators Embedding the original mobile phone chip and enhancing the image function of the chip in the market, Google is not investing in the artificial intelligence industry, developing a specific machine learning chip. Compared with other manufacturers, it is based on the development of ecological cutting, the specific creation of the company Graphcore. Inferential wafer technology is a leader and favors many investors. In the development ecology, Google and Nvidia are relatively complete. Intel strives to catch up. Apple's application software development ecosystem is the most powerful. If you can guide the development of artificial intelligence applications, energy should not be underestimated. Overall, Google, Nvidia, Intel, and Graphcore are competitive conditions for more artificial chips. The development of the chip is currently a competitive situation. Apart from leading manufacturers, they are more active in developing new chips or acquiring new artificial intelligence chip manufacturers. However, many new chip companies, through venture capital, also hope to find the application of AI chips, in order to enter the positive cycle of capital and R&D. In short, the interactive research and development of the chip and the application market support the rapid expansion of the application field. The newly created chip design company also has a chance to find the application direction. It is also worthwhile for the leading manufacturers in the market to apply accurate product strategies.

一、 中文文獻
行政院新聞傳播處. (2018年7月). 台灣的「AI小國大戰略. 擷取自 https://www.ey.gov.tw/Page/5A8A0CB5B41DA11E/50a08776-e33a-4be2-a07c-a6e523f5031b
林守德. (2017). 台大資工系教授林守德：論第三波人工智慧技術革命. 擷取自 科技新報: https://technews.tw/2017/08/07/the-3rd-revolution-wave-of-ai-technology/
張孟凡. (2017). 清華4論文入選ISSCC晶片奧運會. 擷取自 www.nthu.edu.tw: https://www.nthu.edu.tw/hotNews/content/545
陳俊勳. (2013). 擴散創新理論. 國立東華大學.
蔡志宏,劉芳梅,鄭雅嵐. (2018). 2018日本科技創新政策推動與數位經濟創新應用考察團參訪報告. 行政院.
賴志遠. (2018). 國際人工智慧政策推動現況. STPI.
二、 英文文獻
4 AI Trends that will Transform the Telecom Industry in 2019. (2018). 擷取自 Becoming Human: https://becominghuman.ai/4-ai-trends-that-will-transform-the-telecom-industry-in-2019-1bf0d58637cd
AfsharVala. (2018). AI will transform product management. 擷取自 ZDNet: https://www.zdnet.com/article/ai-is-transforming-product-management/
AI System Developer Statistic. (2017). 擷取自 Monster.com.
Ajit Nazre and Rahul Garg. (2015). A Deep Dive in the Venture Landscape of Artificial Intelligence and Machine Learning.
Barbara J. Grosz, Chair, Russ Altman, Eric Horvitz, Alan Mackworth, Tom Mitchell, Deidre Mulligan, Yoav Shoham. (2016年8月). ARTIFICIAL INTELLIGENCE AND LIFE IN 2030. One Hundred Year Study on Artificial Intelligence (AI100).
BayernMacy. (2019年1月). 6 ways AI will revolutionize retail. 擷取自 techrepublic.com: https://www.techrepublic.com/article/6-ways-ai-will-revolutionize-retail/
CoreaFrancesco. (2018). AI Knowledge Map: How To Classify AI Technologies. 擷取自 https://www.kdnuggets.com/2018/08/ai-knowledge-map-classify-ai-technologies.html.
ED FELTEN and TERAH LYONS. (2016年10月12日). The Administration's Report on the Future of Artificial Intelligence. 擷取自 The WHITE HOUSE President BARACK OBAMA: https://obamawhitehouse.archives.gov/blog/2016/10/12/administrations-report-future-artificial-intelligence
Faridah DjellalGalloujFaïz. (2014). The laws of imitation and invention: Gabriel Tarde and the evolutionary. 頁 6-10.
Gartner. (2018). Hype Cycle for Artificial Intelligence, 2018.
GodinBenoît. (2013). Invention, Diffusion and Linear Models of Innovation. Quebec: Project on the Intellectual History of Innovation.
Haris Gacanin, Mark Wagner. (2018). Artificial Intelligence Paradigm for Customer Experience Management in Next-Generation Networks: Challenges and Perspectives. IEEE Network Magazine.
HeinemeyerMax. (2018). The Next Paradigm Shift: AI-Driven Cyber-Attacks. DARKTrace.
MBALib. (2008). 波特五力分析模型. 擷取自 MBALib: https://wiki.mbalib.com/zh-tw/%E6%B3%A2%E7%89%B9%E4%BA%94%E5%8A%9B%E5%88%86%E6%9E%90%E6%A8%A1%E5%9E%8B
McCarthyJohn. (1955). A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence.
MoorJames. (2007). The Dartmouth College Artificial Intelligence Conference: The Next Fifty Years. AI Magazine Volumn 27 Number 4 2007.
PBS (導演). (1992). The Machine That Changed The World [動畫].
PracharSocial. (無日期). artificial-intelligence. 擷取自 https://socialprachar.com/: https://socialprachar.com/
Rahul Kumar, Supradip Baul. (2019). Artificial Intelligence Chip Market Outlook - 2025. Allied Market Research.
ReddyRajD. (1976年4月). Speech Recognition by Machine: A Review". Proceeding of the IEEE, 頁 501-531.
RosenblattFrank. (1957). The Perceptron—A Perceiving and Recognizing Automaton. Cornell Aeronautical Laboratory Report.
SchadtTanjeff. (2018). Artificial Intelligence – The next big growth driver for the semiconductor industry. PwC Strategy.
SelfridgeG.O. (1959). Pandemonium: A paradigm for learning. Proceedings of the Symposium on, 頁 511-531.
SingerPete. (2018年9月). AI Chips: Challenges and Opportunities. 擷取自 https://semimd.com/: https://semimd.com/blog/2018/09/12/ai-chips-challenges-and-opportunities/
(2018). SUMMARY OF THE 2018 WHITE HOUSE SUMMIT ON ARTIFICIAL INTELLIGENCE FOR AMERICAN INDUSTRY. Select Committee on Artificial Intelligence.
Svetiana Sicular, Kenneth Brant. (2018). Hype Cycle for Artificial Intelligence, 2018.
TeichPaul. (2018年8月). DESIGNING CUSTOM CHIPS IN-HOUSE IS THE NEW NORMAL. 擷取自 www.nextplatform.com: https://www.nextplatform.com/2018/08/08/designing-custom-chips-in-house-for-specific-tasks-is-the-new-normal/
WIPO. (2019). WIPO Technology Trends 2019 - Artificial Intelligence.
Yoav Shoham, Raymond Perrault, Erik Brynjolfsson, Jack Clark, James Manyika, Juan Carlos Niebles,. (2018). The AI Index 2018 Annual Report. Human-Centered AI Initiative, Stanford University.
Zheng YouWeiShaojun. (2018). White Paper on AI Chip 2018. Beijin: Beijing Innovation Center for Future Chips (ICFC).