为何英国人不爱搞晶圆厂？

在先前一篇讨论半导体制造投资的文章中，有网友问到：“为何英国半导体业者都放弃了晶圆厂，最后演变成像 ARM 那样的芯片/IP设计公司？”，以及：“为何在英国传统的死对头国家法国，晶圆厂反而一直存在(虽然也不算兴盛)？”以下笔者尝试响应这些问题。(编按：本文作者为英国籍) 详细的原因是很复杂的，但讲得太笼统又恐怕有缺失；笔者将以一个1984年出道的中立观察者角度来谈英国电子产业。英国半导体晶圆厂不兴盛的主要原因，是第二次世界大战让美国成为世界经济强权，在1950到1960年代初对英国国力产生严重冲击，影响甚至持续到今日。另一个原因则是，与科学领域相较，英国的人文艺术发展历史更悠久。 我认为症结点在于科技并不足以代表英国的政治与经济成就，但也许这只是我的个人观点。在遥远的过去，英国企业的拥有者与经营团队，通常都没有科技背景；诸如ARM、Wolfson、CSR、Vodafone等数家表现出色的公司，都是比较近期才诞生的例外。 回溯到1960年代，英国的科技公司管理阶层都把工程师当成奴工般对待，不需要也不应该支付太高的薪资；这种“将就凑合着用(make do and mend)”的心态是第二次世界大战的遗毒，而且英国许多电子公司还是以早期对电子业有所需求的军事防卫应用为主。这类公司通常会有以下缺点： 1. 不具备国际竞争力；
2. 无法支应电子产业所需的、不断升高的营运成本；
3. 吝啬而且眼界狭隘。 在1960与1970年代，当面临“该花费数百万英镑竞逐薄膜单石集成电路(thin-film monolithic integrated circuits)？”，还是“只花数千英镑来生产厚膜混合IC与PCB？”的抉择时，英国大大小小的电子业高层都选择后者，只看到成本没看到两者间的其它差异性。 在那个时候，美国与英国的电子产业在财力、抱负与创业精神方面的差距，以及美国可提供高薪、英国本地薪资微薄等等条件，造成了英国出现严重的人才外流。例如曾任职于摩托罗拉(Motorola)与快捷半导体(Fairchild)、并创立LSI Logic的资深工程师Wilf Corrigan，就是英国利物浦码头工人之子。 人才外流进一步加深了英美电子产业之间的差距，间接让美国发展出顶尖的芯片制造产业，也让英国在这方面更为落后。我记得在1984年参观英国的一座晶圆厂，它位于一座几十年历史的老建筑，有破窗户与堆置在角落的化学品圆桶，看起来就像是骯脏的修车厂。 英国电子业者总部通常座落于空间狭窄的维多利亚时代建筑，对于布建需要独立厂房、设备成本动辄数万英镑的制程兴趣缺缺。因此大多数的英国晶圆厂，其实是在某段时间外商投资者为了因应可能兴起的“欧洲堡垒(Fortress Europe；编按：成立欧盟的初始构想)”，所以预先在欧洲寻找据点。 到了1980与1990年代，面临要花数百万英镑设计ASIC，或是进行几乎没有非重复性工程费用(NRE)的FPGA编程的抉择时，英国业者通常选择后者。在1990年，就算有Apple、VLSI Technology、Acorn与一家日本投资银行的支持，创建一家处理器公司几乎不需要资金；这种模式的必要条件，是建立智财(IP)业务模式，也是ARM迄今奉行不悖的。 看到这里，你有发现任何极度吝啬或是资金匮乏的因素在其中吗？那到底为什幺目前在法国仍然拥有当地业者自有自营的、接近顶尖的晶圆厂，但英国境内晶圆厂那幺少？嗯…撒切尔夫人与意法半导体(ST)执行长Pasquale Pistorio 是两个原因。 在撒切尔夫人担任首相的1979年到1990年之间，英国走右派路线，废除了给予GEC、Plessey、Ferranti与Marconi等半导体业者的补助，虽然GEC-Plessey Semiconductors曾一度成为全英国最大的芯片厂商，但显然在后撒切尔时代的英国，半导体产业已经式微。 在此同时，曾在美国半导体大厂Motorola工作的Pistorio，满怀热情地认为半导体产业会带来财富；他与中间边左派的法国与意大利政府合作，这些国家采纳政府干预策略，给予Philips、Siemens与ST等欧洲大厂不少支持。 如同英国曾经历过的，全球化效应现在看来让美国开始走下坡；从某部分看来，风险资金(如前面所提到的种种原因，英国几乎不曾涉入这种投资)正转向其它领域，目标由信息技术所需的半导体产业，转向发电、节省能源与诉求环境永续性的半导体。 总之，没错，在放弃半导体制造这件事上，英国确实领先美国；他们也走在台湾、中国的前面。但这到底是好、是坏还是不可避免的趋势呢？这又是另一个议题了。 编译：Judith Cheng 本文授权编译自EE Times，版权所有，谢绝转载 参考英文原文： Why the British got out of fabs，by Peter Clarke

相关阅读：
• iSuppli：2015年12寸晶圆产量将增近一倍
• 智能手机、平板电脑带动下半年全球晶圆代工产值增长12.3%
• 解析无晶圆半导体公司MEMS收入构成及代工关系vx6esmc

{pagination} Why the British got out of fabs Peter Clarke In response to a request from a participant in the forum discussion below When the smart money got into/out of manufacturing I have tried to answer the questions: "Why the British had to get out of fabs and end up just with design/IP a la ARM ?" and "Why fabs still survive (if not exactly thrive) in the UK's traditional rival France?" The detailed reasons are complex, generalizations are usually faulty but I will have a go at putting down my perspective as an observer of the electronics industry from the U.K. since 1984. A primary reason that fabs failed to thrive in the U.K. is that while the Second World War helped create the U.S. as a global economic superpower it more or less bankrupted the U.K. with implications that were heavy in the 1950s and early 1960s and that continue to this day. A second reason is a long-established arts and humanities versus science division in U.K. society. I think it is still the case that science and technology are not sufficiently well represented amongst the political and wealthy establishment in the U.K. But maybe that is just the science graduate in me talking. In the distant past the owners and managers of UK companies usually had a non-scientific background. The likes of ARM, Wolfson, CSR, Vodafone and several others are now glorious but relatively recent exceptions. Back in 1960s U.K. technology-based company management treated engineers as willing serfs who did not need and should not be given too much money. It was a "make do and mend" mentality left over from the Second World War, and many electronics companies continued to be closely aligned to military interests and defense was the early application for electronics. These companies often could not comprehend or cope with: 1) international competition 2) the continuous exponential increase in the cost of participation in electronics Parsimonious and parochial When faced in the 1960s and 1970s with the choice between spending millions of pounds to compete in thin-film monolithic integrated circuits and thousands of pounds to make thick-film hybrids and PCBs, managements up and down the U.K. chose to do the latter not perceiving any difference but cost. The difference between the wealth, ambition and entrepreneurial spirit abroad in the United States, and the high salaries paid there, and the meager existence in the United Kingdom at this time, gave rise to a notorious exodus of talent, known as the "brain drain." Included in this were such people as Wilf Corrigan, the son of a Liverpool docker who worked at Motorola and Fairchild and founded LSI Logic. This migration of talent tended to polarize the situation further and help the United States pioneer chip production for higher industrial and professional volumes and the U.K. to fall behind. I remember visiting a wafer fab in the U.K. in 1984 that had been housed in a building that was decades old, had broken windows and drums of chemicals out the back stored on a concrete standing. It looked like and was as grubby as an automobile bodyshop. U.K. companies were often on space-constrained Victorian sites and few had any appetite for a manufacturing process that required its own building and equipment that cost hundreds of thousands of pounds. Therefore most of the wafer fabs that were built in U.K. were done so by inward investors sometimes seeking to gain a European location as a hedge against the possibility of an emerging "Fortress Europe." When faced in the 1980s and 1990s with a choice between spending millions of pounds to design ASICs or almost no non-recurring expense (NRE) to program FPGAs U.K. companies often chose to design using FPGAs. In 1990 when creating a processor company the option was taken to do it on almost no money, albeit backed by Apple, VLSI Technology, Acorn and a Japanese investment bank. This approach necessitated the creation of the intellectual property business model that has been pursued ever since by ARM. Are we detecting a parsimonious capital-starved theme here? one moved to the right, one didn't And why do indigenously-owned and operated fabs continue at close to the leading edge in France but there are so few fabs in UK? Well Margaret Thatcher and Pasquale Pistorio are two reasons. The move to the right in the U.K. under Thatcher's tenure as prime minister from 1979 to 1990 swept away the idea of subsidies for the likes of GEC, Plessey, Ferranti and Marconi and although GEC-Plessey Semiconductors Ltd. became a de facto national champion chip company for a brief period the concept was clearly out of fashion in post-Thatcher UK. Meanwhile Pasquale Pistorio believed passionately in the power of semiconductors to create wealth, had worked inside a U.S. pioneer of the semiconductor industry, Motorola, and was working with the governments of France and Italy that were left-of-center, believed in state intervention and supported European champions such as Philips, Siemens and STMicroelectronics. Globalization does appear to be driving the U.S. down the same paths already taken by the U.K. But partly what we are seeing is that venture capital (which the UK never had to any degree for reasons mentioned at the top of the article) is turning to other things. It is turning away from semiconductors for information technology, and towards other things, such as semiconductors for power generation, for energy-efficiency and sustainability. But, yes, the U.K. does lead the U.S. in terms of getting out of manufacturing. And both are ahead of Taiwan and China. But whether that, is a good, bad or inevitable, is another matter of opinion.